<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by PV Prasanth on Medium]]></title>
        <description><![CDATA[Stories by PV Prasanth on Medium]]></description>
        <link>https://medium.com/@pvprasanth474?source=rss-f882d8a1979f------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*lOAct_ax8pOmJiHT7-mcDw@2x.jpeg</url>
            <title>Stories by PV Prasanth on Medium</title>
            <link>https://medium.com/@pvprasanth474?source=rss-f882d8a1979f------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 02:12:09 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@pvprasanth474/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Free AI Certifications by Anthropic (Claude) — A Complete Guide]]></title>
            <link>https://medium.com/@pvprasanth474/free-ai-certifications-by-anthropic-claude-a-complete-guide-a2b94bd82a34?source=rss-f882d8a1979f------2</link>
            <guid isPermaLink="false">https://medium.com/p/a2b94bd82a34</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[calude-code]]></category>
            <category><![CDATA[anthropic-claude]]></category>
            <category><![CDATA[claude-ai]]></category>
            <dc:creator><![CDATA[PV Prasanth]]></dc:creator>
            <pubDate>Wed, 08 Apr 2026 03:59:42 GMT</pubDate>
            <atom:updated>2026-04-08T03:59:42.865Z</atom:updated>
            <content:encoded><![CDATA[<p>No paywalls. No subscriptions. No upsells. Just free courses from the team that builds Claude.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*O7Ko87aMmhCVzmL4nOswGQ.png" /><figcaption>Anthropic Courses</figcaption></figure><p>Here’s an honest breakdown of what they offer and where to start.</p><p><strong>Why This Is Worth Paying Attention To</strong></p><p>Most free AI courses follow the same playbook — a few videos, surface-level tips, then a pitch for the paid version.</p><p>Anthropic’s courses are different. They’re built by the people actually working on AI, and it shows. The focus isn’t just on <em>using</em> AI. It’s on understanding it well enough to use it consistently and correctly.</p><p>That’s a rare thing to find for free.</p><p><strong>Who It’s For</strong></p><p>More people than you’d think.</p><ul><li><strong>Developers</strong> who want to build real AI applications</li><li><strong>Students</strong> who want to study smarter</li><li><strong>Educators</strong> adapting to AI in the classroom</li><li><strong>Professionals</strong> who want AI to actually improve their work</li><li><strong>Nonprofits</strong> trying to do more with less</li></ul><p>There are 13 courses total, organized into three tracks. Let’s keep it simple.</p><p><strong>The Three Tracks</strong></p><p><strong>AI Fluency </strong>— Learn how to think and communicate with AI. This is where everyone should start.</p><p><strong>Developer and Technical </strong>— Build apps, use APIs, create AI workflows. For people who want to go from user to builder.</p><p><strong>Specialized</strong> — Focused courses for students, educators, nonprofits, and teams on AWS or Google Cloud.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*W4tzBBzrcFvWdLyVkEv1Aw.png" /></figure><p><strong>The 13 Courses, Simply Explained</strong></p><p><strong>1. Claude 101</strong></p><p><em>Best for: Everyone</em></p><p>The starting point. Real examples, practical use cases, no fluff. You’ll finish this knowing how to actually use Claude in your daily work — not just in theory.</p><p><strong>2. AI Fluency: Framework and Foundations</strong></p><p><em>Best for: Everyone</em></p><p>This is the most underrated course here. It teaches you how to prompt well, how to collaborate with AI, and — just as importantly — where AI goes wrong. Most people skip this. Don’t.</p><p><strong>3. Claude Code in Action</strong></p><p><em>Best for: Developers and technical teams</em></p><p>How to use AI inside real coding workflows. Saves hours on repetitive tasks. Practical from the first lesson.</p><p><strong>4. Building with the Claude API</strong></p><p><em>Best for: Developers who want to build products</em></p><p>Goes from basics to building real, working applications. API calls, streaming, tool use — all covered in a logical sequence.</p><p><strong>5. Introduction to Model Context Protocol (MCP)</strong></p><p><em>Best for: Developers building AI systems</em></p><p>MCP is how AI connects to the real world — external tools, data sources, services. It’s becoming a standard fast. Getting familiar with it now puts you ahead.</p><p><strong>6. MCP Advanced Topics</strong></p><p><em>Best for: Developers in production environments</em></p><p>Production architecture, file systems, notifications. For those going deep on building reliable AI systems.</p><p><strong>7. AI Fluency for Students</strong></p><p><em>Best for: Students</em></p><p>Shows you how to use AI to learn better, not just produce outputs faster. That’s the skill that actually compounds over time.</p><p><strong>8. AI Fluency for Educators</strong></p><p><em>Best for: Teachers and instructors</em></p><p>How to bring AI into your classroom thoughtfully. Practical workflows, not just theory.</p><p><strong>9. AI Fluency for Nonprofits</strong></p><p><em>Best for: Nonprofit teams</em></p><p>How to use AI to extend your reach without extending your budget. Specific and useful.</p><p><strong>10. Claude with Amazon Bedrock</strong></p><p><em>Best for: Teams on AWS</em></p><p>Deploying Claude in enterprise AWS environments. Bridges the gap between the model and your infrastructure.</p><p><strong>11. Claude with Google Cloud Vertex AI</strong></p><p><em>Best for: Teams on Google Cloud</em></p><p>Same idea, for Google Cloud users.</p><p><strong>12. Teaching AI Fluency</strong></p><p><em>Best for: L&amp;D professionals and team leads</em></p><p>How to build AI literacy in others. If your job involves training people, this one is for you.</p><p><strong>13. Introduction to Agent Skills</strong></p><p><em>Best for: Anyone curious about where AI is heading</em></p><p>AI agents are the future — systems that act on your behalf, not just answer questions. This course shows you how to start building them.</p><p><strong>How Long Will It Take?</strong></p><p>Most courses are 1–4 hours. You can realistically finish two or three in a weekend.</p><p><strong>How to Get Certified</strong></p><p>1. Click on <a href="https://anthropic.skilljar.com/"><strong>Anthropic Academy on Skilljar</strong></a></p><p>2. Create a free account</p><p>3. Pick a course — listed at $0</p><p>4. Complete the lessons</p><p>5. Download your certificate</p><p>That’s it.</p><p><strong>Where to Start</strong></p><p>Not sure which course to take first? Here’s a simple path:</p><p><strong>If you’re new to AI:</strong></p><p>Claude 101 → AI Fluency: Framework and Foundations</p><p><strong>If you want to build:</strong></p><p>Building with the Claude API → Introduction to MCP</p><p><strong>If you’re in education or nonprofits:</strong></p><p>Go straight to your specialized track after Claude 101</p><p><strong>Are the Certificates Worth It?</strong></p><p>Honestly — the certificate itself is a bonus.</p><p>The real value is the knowledge. AI is being factored into hiring, strategy, and operations everywhere. Understanding how it actually works, where it fails, and how to use it well is becoming a baseline professional skill.</p><p>These courses give you that foundation. For free.</p><p><strong>Final Thought</strong></p><p>Anthropic didn’t have to make these free. They did anyway.</p><p>Thirteen courses. Structured learning. Real certifications. From the people building one of the leading AI systems in the world.</p><p>Whether you’re just getting started or want to go deep — there’s something here for you.</p><p><em>Which course are you planning to take first? Drop it in the comments — I’m curious what’s most useful for people in different roles.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a2b94bd82a34" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Agentic AI: Understanding the 6 Building Blocks]]></title>
            <link>https://medium.com/@pvprasanth474/agentic-ai-understanding-the-6-building-blocks-6ff1dc819cfa?source=rss-f882d8a1979f------2</link>
            <guid isPermaLink="false">https://medium.com/p/6ff1dc819cfa</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[ai-agents-in-action]]></category>
            <dc:creator><![CDATA[PV Prasanth]]></dc:creator>
            <pubDate>Fri, 06 Mar 2026 07:39:34 GMT</pubDate>
            <atom:updated>2026-03-06T07:39:34.622Z</atom:updated>
            <content:encoded><![CDATA[<p>In our previous articles, we explored <a href="https://medium.com/@pvprasanth474/what-is-agentic-ai-c53121adb71a">what Agentic AI</a> is and <a href="https://medium.com/@pvprasanth474/what-powers-an-agentic-workflow-7f3a70b9325b">what powers an agentic workflow </a>— covering AI agents, LLMs, tools, feedback loops, multi-agent collaboration, and more.</p><p>Now let’s go deeper — <strong>what are the core modules inside every agentic system?</strong></p><p>Every Agentic AI system is built from <strong>6 fundamental building blocks</strong>:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*t2l03gFGyN6RdECZ1B3W4A.png" /><figcaption>Building Blocks of Agentic AI</figcaption></figure><p>Think of them like the organs of a human body:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/396/1*mO1IyNWExSVQKPZOnMwvSw.png" /></figure><p>Let’s break each one down.</p><h3>1.Perception Module — The Senses</h3><p>The agent’s sensory system. It <strong>gathers and interprets data</strong> from the environment — text, images, APIs, databases, sensors, web pages.</p><p><strong>Example — Healthcare Monitoring Agent:</strong></p><pre>INPUT SOURCES:<br>├── Heart rate monitor     → 98 bpm        → tagged: normal<br>├── Blood pressure sensor  → 145/95 mmHg   → tagged: elevated<br>├── Temperature sensor     → 38.4°C        → tagged: mild fever<br>├── Patient medical record → History of hypertension<br>└── Nurse&#39;s notes          → &quot;Patient has a headache&quot;<br><br>PERCEPTION OUTPUT (structured data passed forward):<br>→ { heart_rate: &quot;normal&quot;, bp: &quot;elevated&quot;, temp: &quot;mild_fever&quot;,<br>    history: [&quot;hypertension&quot;], symptoms: [&quot;headache&quot;] }</pre><p>The Perception Module doesn’t decide what to do — it <strong>senses, organizes, and passes context</strong> to the Cognitive Module.</p><h3>2. Cognitive Module — The Brain</h3><p>This is where the agent <strong>thinks, reasons, plans, and reflects</strong>. Powered by LLMs and structured reasoning techniques:</p><ul><li><strong>Chain of Thought</strong> — think step by step</li><li><strong>Subgoal Decomposition</strong> — break big goals into small tasks</li><li><strong>Self-Reflection</strong> — evaluate and improve own output</li></ul><p><strong>Example — Travel Planning Agent:</strong></p><pre>User: &quot;Plan a 5-day Japan trip for 4 people under $5,000.&quot;<br><br>REASONING:<br>&quot;Flights at $1,200/person = $4,800... that&#39;s the whole budget!<br> Let me find budget flights at ~$700/person = $2,800.<br> Remaining $2,200 → hotel $500, food $300, transport $400, <br> activities $200. Total: $4,200 ✅ Under budget!&quot;<br><br>SELF-REFLECTION:<br>&quot;Wait — I don&#39;t know where they&#39;re flying from. <br> I&#39;ll note that flight prices vary by departure city.&quot;</pre><p>The Cognitive Module reasons through the problem, catches its own mistakes, and makes smart decisions.</p><h3>3. Action Module — The Hands and Feet</h3><p>Where decisions become <strong>real actions</strong> — calling APIs, sending emails, executing code, querying databases, browsing the web.</p><p><strong>Example — Customer Service Agent:</strong></p><pre>Customer: &quot;My laptop order hasn&#39;t shipped in 3 days!&quot;<br><br>ACTION MODULE EXECUTES:<br>Step 1: get_order(&quot;C-9921&quot;) → Order found, shipping overdue<br>Step 2: check_inventory(&quot;ProBook-15&quot;) → Out of stock!<br>Step 3: check_inventory(&quot;ProBook-17-Pro&quot;) → 12 in stock<br>Step 4: get_loyalty_tier(&quot;C-9921&quot;) → Gold member, $200 credit<br>Step 5: send_email(&quot;Free upgrade to ProBook 17 Pro!&quot;)<br>Step 6: update_order(new_item=&quot;ProBook-17-Pro&quot;, shipping=&quot;express&quot;)</pre><p>6 tool calls, fully autonomous — turning a complaint into a delightful experience.</p><h3>4. Learning Module — Experience &amp; Growth</h3><p>Gives the agent <strong>memory</strong> — both short-term (current session) and long-term (across sessions) — so it improves over time.</p><p><strong>Example — AI Coding Tutor:</strong></p><pre>Monday:    Student asks about decorators → agent gives complex answer<br>           Student says &quot;Too complicated!&quot; <br>           📝 Stores: prefers simple explanations, beginner level<br><br>Wednesday: Student asks about classes<br>           🧠 Recalls: this student likes analogies<br>           🤖 &quot;Think of a class like a cookie cutter 🍪 — <br>              it defines the shape, each cookie is an object!&quot;<br>           Student: &quot;Makes sense! 👍&quot;<br><br>Friday:    Student asks about databases<br>           🧠 Recalls: beginner but progressing<br>           🤖 Gives slightly deeper explanation with SQLAlchemy</pre><p>The agent adapted its teaching style across three sessions — <strong>without being reprogrammed</strong>.</p><h3>5. Collaboration Module — Teamwork</h3><p>Enables <strong>multiple specialized agents to work together</strong>, sharing context through shared memory.</p><p><strong>Example — Content Creation Team:</strong></p><pre>User: &quot;Write a researched blog post about remote work.&quot;<br><br>PROJECT MANAGER AGENT coordinates:<br><br>1. 🔍 Research Agent → finds stats from McKinsey, Stanford studies<br>   → stores in SHARED MEMORY<br><br>2. ✍️ Writer Agent → reads research, drafts 1,500-word article<br>   → stores in SHARED MEMORY<br><br>3. 📝 Editor Agent → reviews: &quot;Fix citation in para 3, <br>   strengthen the hook, add SEO keyword&quot;<br>   → sends feedback to Writer<br><br>4. ✍️ Writer Agent → revises based on feedback<br><br>5. 📱 Social Media Agent → creates Twitter thread + LinkedIn post</pre><p>Five agents, each with a specialty, collaborating like a digital content team.</p><h3>6. Security Module — The Immune System 🛡️</h3><p>The <strong>protective layer</strong> that keeps the agent safe, ethical, and within boundaries. Without it, an autonomous agent with real-world access is a risk.</p><p>Four layers of protection:</p><ul><li><strong>Input Security</strong> — blocks prompt injection, validates requests</li><li><strong>Execution Security</strong> — action limits, budget caps, human approval for risky operations</li><li><strong>Output Security</strong> — filters harmful content, redacts sensitive data</li><li><strong>Audit &amp; Observability</strong> — logs every action, tracks costs, ensures compliance</li></ul><p><strong>Example — Trading Agent:</strong></p><pre>🤖 Agent: &quot;I recommend selling 100 shares ($4,500)&quot;<br><br>🛡️ Security: Trade exceeds $2,000 → HUMAN APPROVAL REQUIRED<br><br>📱 User gets notification:<br>   &quot;SELL: 100 × Company X = $4,500&quot;<br>   [✅ Approve]  [❌ Reject]  [✏️ Modify]<br><br>User: [✅ Approve] → Trade executed → Full audit log created</pre><p>The agent acts autonomously for small decisions but <strong>escalates to the human</strong> for high-risk ones.</p><p>These 6 modules operate in a continuous cycle:</p><pre>SENSE → REASON → PLAN → COORDINATE → ACT → LEARN &amp; ADAPT<br>  ↑                                                    │<br>  └────────── feedback loop ───────────────────────────┘<br><br>  🛡️ Security Module — active across ALL steps</pre><p>These 6 building blocks are the <strong>architecture behind every agentic AI system</strong>. Understanding them is the foundation for designing, building, and evaluating AI agents in the real world.</p><p>🙌 <strong>If You Found This Helpful…</strong></p><p>👉 Give this article a clap — it helps others discover it 👉 Share it with anyone curious about Agentic AI 👉 Follow me on Medium for more in this Agentic AI series</p><p>Stay tuned — more coming soon! 🚀</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6ff1dc819cfa" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Powers an Agentic Workflow?]]></title>
            <link>https://medium.com/@pvprasanth474/what-powers-an-agentic-workflow-7f3a70b9325b?source=rss-f882d8a1979f------2</link>
            <guid isPermaLink="false">https://medium.com/p/7f3a70b9325b</guid>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[agentic-workflow]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[PV Prasanth]]></dc:creator>
            <pubDate>Wed, 11 Feb 2026 12:10:04 GMT</pubDate>
            <atom:updated>2026-02-11T12:10:04.710Z</atom:updated>
            <content:encoded><![CDATA[<p>Most people think <strong>agentic AI</strong> is just a smarter chatbot. It’s not.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pMQo73GKaGUaHMf5v-z8lA.png" /><figcaption>Agentic workflow</figcaption></figure><p>Today’s AI systems can <strong>take actions</strong>, <strong>use tools</strong>, <strong>reason</strong>, <strong>collaborate</strong>, and even <strong>self‑correct</strong> — all without constant human supervision.</p><p>This new era is called <strong>Agentic AI</strong>.</p><p>An agentic system is closer to a <strong>digital worker</strong> — one that can plan, act, check its own work, and keep going until a goal is achieved.</p><p>But what exactly <em>powers</em> an agentic workflow?</p><h3>What Is an Agentic Workflow?</h3><p>An <em>agentic workflow</em> is a process where AI:</p><ol><li><strong>Observes</strong> a situation</li><li><strong>Thinks</strong> about what to do</li><li><strong>Acts</strong> using tools</li><li><strong>Learns</strong> from feedback</li><li><strong>Repeats</strong></li></ol><p>It’s like giving AI a goal and letting it figure out the steps on its own.</p><p>To make this possible, 7 major components work together.</p><h3>1. AI Agents — The Autonomous Workers</h3><p>AI agents are software systems designed to operate <strong>independently</strong>.<br> They can take actions, make decisions, and complete tasks based on a goal.</p><p><strong>Example:</strong></p><p>Think of an AI agent that:</p><ul><li>Reads your emails</li><li>Sorts them</li><li>Replies automatically</li><li>Reminds you about urgent messages</li></ul><p>The agent is doing work — not just generating text.</p><h3>2. Large Language Models (LLMs) — The Brain Behind the Agent</h3><p>LLMs like GPT‑4, GPT‑5, Claude, or Llama provide the <strong>intelligence</strong> required for agents.</p><p>They help agents:</p><ul><li>Understand natural language</li><li>Reason about tasks</li><li>Break down problems</li><li>Generate solutions</li></ul><p>Without LLMs, agents wouldn’t be able to “think”.</p><h3>3. Tools — The Agent’s Hands and Superpowers</h3><p>A <strong>tool</strong> is anything the agent uses to take an action <strong>outside the LLM</strong>.<br> Think of tools as the <strong>hands, eyes, calculator, and apps</strong> the agent uses to get real work done.</p><p>Without tools → an agent can only <em>talk</em>.<br> With tools → an agent can <em>act</em>.</p><p><strong>Example:</strong><br> You ask: <em>“Tell me today’s weather in Bengaluru.”</em><br> The agent uses the <strong>search tool</strong> to fetch live weather data.</p><h3>4. Feedback Mechanisms — Self‑Correction Loop</h3><p>Feedback is how agents stay accurate and improve outcomes.</p><p>Agents can:</p><ul><li>Review their own results</li><li>Detect mistakes</li><li>Retry actions</li><li>Request clarification</li><li>Adjust their approach</li></ul><p>This is like giving AI a built‑in quality assurance system.</p><p><strong>Example:</strong></p><p>An agent writes code → tests fail → agent fixes bugs automatically.</p><h3>5. Prompt Engineering — The Rules and Instructions</h3><p>Creating clear instructions that guide how the agent behaves, reasons, and acts.</p><p>Prompts define the <strong>rules of the agent</strong>, like:</p><ul><li>“Always check your reasoning.”</li><li>“Before acting, explain the plan.”</li></ul><p><strong>Example:</strong><br> Prompt: <em>“You are my daily task manager. Each morning, read my calendar, summarize tasks, and suggest priorities.”</em></p><h3>6.Multi‑Agent Collaboration — AI Teams Working Together</h3><p>Complex workflows often need more than one agent.</p><p>Example team setup:</p><ul><li>Planner Agent → decides the steps</li><li>Research Agent → gathers info</li><li>Developer Agent → writes code</li><li>Tester Agent → validates output</li></ul><p>Each agent specializes in one task.<br> Together, they can solve problems no single agent could.</p><h3>7.Integrations — Connecting AI to Your World</h3><p>Connecting agents with your existing systems — Outlook, SharePoint, CRMs, GitHub, Databases, Jira, cloud services, etc.</p><p>Integrations allow agents to use your existing systems:</p><ul><li>Outlook</li><li>SharePoint</li><li>GitHub</li><li>Jira</li><li>Databases</li><li>CRMs</li><li>Cloud services</li></ul><p>Without integrations, agentic AI would stay isolated.<br> With integrations, it becomes a powerful part of your workflow.</p><h3>Conclusion</h3><p>Agentic AI represents the next major leap in how we build and interact with intelligent systems. Instead of simply generating text, AI agents can now <em>think, act, learn, and collaborate</em> using tools, feedback loops, and integrations.</p><p>By combining LLMs, prompt engineering, tool usage, multi‑agent teamwork, and real‑world system connections, agentic workflows enable AI to become a true digital teammate — one that can automate tasks, reason through complex processes, and deliver outcomes with minimal human intervention.</p><p>Whether you’re an engineer, product builder, business leader, or AI enthusiast, understanding these foundational components is the key to unlocking the future of autonomous AI systems.</p><p>The bottom line:<br> 👉 <strong>Agents aren’t just the next step in AI — they’re the beginning of a new era of intelligent automation.</strong></p><h4><strong>🙌 If You Found This Helpful…</strong></h4><p>If this article helped you understand how agentic workflows work — and sparked your curiosity about the future of autonomous intelligence — I’d love your support:</p><p>👉 <strong>Give this article a clap (or many!)</strong> — it helps more readers discover it<br> 👉 <strong>Share it</strong> with friends, colleagues, or anyone exploring AI and automation<br> 👉 <strong>Follow me on Medium</strong> for more deep‑dives into AI agents, reasoning systems, autonomy, and the next wave of intelligent tools</p><p>Your support motivates me to keep creating high‑quality, beginner‑friendly AI content.</p><p>🚀 <strong>Stay tuned — the next article in this Agentic AI series is coming soon!</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7f3a70b9325b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is Agentic AI?]]></title>
            <link>https://medium.com/@pvprasanth474/what-is-agentic-ai-c53121adb71a?source=rss-f882d8a1979f------2</link>
            <guid isPermaLink="false">https://medium.com/p/c53121adb71a</guid>
            <category><![CDATA[agentic-ai-architecture]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[generative-ai-tools]]></category>
            <dc:creator><![CDATA[PV Prasanth]]></dc:creator>
            <pubDate>Tue, 03 Feb 2026 03:14:51 GMT</pubDate>
            <atom:updated>2026-02-03T03:14:51.212Z</atom:updated>
            <content:encoded><![CDATA[<p>Artificial Intelligence has evolved rapidly over the past few years.<br> We’ve moved from systems that simply <em>respond</em>…<br> to systems that can <em>reason</em>, <em>plan</em>, <em>act</em>, and <em>self‑improve</em>.</p><p>Welcome to the world of <strong>Agentic AI</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/932/1*tQY1EM1aikIgEbnnmXrC8Q.png" /><figcaption>Agentic AI Example</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/932/1*pXF8VWKjjbGSxah30z3a2A.png" /><figcaption>Agentic AI Example</figcaption></figure><p>Agentic AI systems operate autonomously, perceive their environment, making decisions, reasoning, and acting proactively to achieve goals in dynamic environments — continuously and without constant human supervision.</p><h3>Key Characteristics of Agentic AI:</h3><h4>1. Autonomy:</h4><p>The AI can work <strong>on its own</strong>, without being told every step.</p><p><strong>Simple Example</strong>:</p><p>Imagine you tell a helpful assistant:<br> “Clean my room.”</p><p>A <em>non‑autonomous</em> helper would ask every few seconds:<br> “Should I pick up the socks?”<br> “Should I fold the clothes?”<br> “Should I make the bed?”<br> “Should I sweep?”</p><p>An <strong>autonomous</strong> helper would:</p><ul><li>Look at the room</li><li>Decide what needs to be done</li><li>Take actions</li><li>Finish the task</li><li>Tell you it’s done</li></ul><p><strong>Agentic AI example:</strong></p><p>Tell an AI: “Summarize this 20‑page report and send a structured email to my team.”<br> It will:</p><ul><li>Read the report</li><li>Find key points</li><li>Write the summary</li><li>Draft the email</li><li>Format it</li><li>Send it (if allowed)</li></ul><p>All <strong>on its own</strong>, without you prompting each step.</p><h4>2. Goal-Oriented Behavior</h4><p>Agentic AI pursues objectives, not just tasks.</p><p><strong>Simple Example</strong>:</p><p>Ask: “Plan a birthday party for my daughter.”</p><p>A goal-oriented system will:</p><ul><li>Decide theme options</li><li>Check your budget</li><li>Find party venues</li><li>Compare prices</li><li>Create a checklist</li><li>Book what’s needed</li></ul><p>It pursues the <em>outcome</em>, not just gives a suggestion.</p><p><strong>Agentic AI example:</strong></p><p>A research agent given the goal:<br> “Find the top 5 competitors for our product and compare features.”<br> It will search, read, organize, and deliver a full report — not just respond once.</p><h4>3. Adaptability</h4><p>The AI can adjust when things change.</p><p><strong>Example:</strong></p><p>If you tell a delivery robot to take a package to your door and your dog blocks the hallway, an <strong>adaptive</strong> robot will:</p><ul><li>Change its path</li><li>Wait</li><li>Or find another route</li></ul><p>A <em>non‑adaptive</em> system would just stop and say “ERROR.”</p><p><strong>Agentic AI example:</strong></p><p>You tell an agent to send an email, but the email server is down.<br> A truly adaptive agent will:</p><ul><li>Try again later</li><li>Inform you</li><li>Or choose another channel (Slack, Teams)</li></ul><p>It <strong>changes strategy</strong> when something goes wrong.</p><h4>4. Contextual Awareness</h4><p>Able to understand the context</p><p><strong>Simple example:</strong></p><p>If you say “Turn on the lights” in the morning, the assistant understands:</p><ul><li>It’s daytime → maybe blinds should be opened instead.</li></ul><p>If you say it at night → it turns on the lamp.</p><p><strong>Agentic AI example:</strong></p><p>If you tell an AI:<br> “Prepare a report for tomorrow’s meeting.”<br> It notices:</p><ul><li>Tomorrow is Monday</li><li>Meeting starts at 10 AM</li><li>You like your reports in PPT format</li><li>Last time you wanted charts</li></ul><p>So it automatically generates a PPT with charts.</p><p>That’s contextual awareness.</p><h4>5. Action-Oriented</h4><p>Turns decisions into actions</p><p><strong>Simple example:</strong></p><p>A regular AI:<br> “You have 5 meetings today.”</p><p>An action‑oriented AI:</p><ul><li>Reschedules conflicts</li><li>Blocks focus time</li><li>Sends reminders</li><li>Joins virtual meetings automatically</li><li>Takes notes</li></ul><p>It performs <strong>actions</strong>, not just gives answers.</p><h4>6. Self-Improving</h4><p>Learns and improves over time.</p><p><strong>Simple example:</strong></p><p>A child solves a puzzle, realizes a piece doesn’t fit, and tries a different piece.</p><p>A <strong>self-improving agent</strong>:</p><ul><li>Checks its result</li><li>Notices mistakes</li><li>Fixes them</li><li>Learns for next time</li></ul><h4>7. Collaboration</h4><p>Able to coordinate with other agents</p><p><strong>Simple example:</strong></p><p>A house-building team:</p><ul><li>One person designs</li><li>One person buys materials</li><li>One person builds</li><li>One person inspects</li></ul><p>They talk and coordinate.</p><p><strong>Agentic AI example:</strong></p><p>A team of AI agents:</p><ul><li><strong>Research Agent</strong> gathers information</li><li><strong>Writer Agent</strong> produces the content</li><li><strong>Reviewer Agent</strong> checks for accuracy</li><li><strong>Designer Agent</strong> makes visuals</li></ul><p>They collaborate like a team of digital coworkers.</p><h4>8. Tool Utilization</h4><p>Uses external tools — Apps, browsers, APIs, code, calculators, databases.</p><p><strong>Simple example:</strong></p><p>If you tell a kid:<br> “Find the meaning of this word,”<br> the kid knows to grab a dictionary.</p><p><strong>Agentic AI example:</strong></p><p>When asked:<br> “Give me the latest stock price of Apple,”<br> the agent will:</p><ul><li>Search the web</li><li>Use a finance API</li><li>Fetch real numbers</li><li>Calculate trends</li></ul><p>Not just guess — it <strong>uses tools</strong> to get real data.</p><h3>Conclusion</h3><p>Agentic AI represents the next major leap in artificial intelligence.<br> It moves beyond passive responses toward <em>autonomous intelligence</em> capable of:</p><ul><li>Acting independently</li><li>Handling complex goals</li><li>Collaborating across tasks</li><li>Learning and improving endlessly</li><li>Integrating deeply with tools, apps, and workflows</li></ul><p>We are heading into a world where AI won’t just answer questions…<br> It will <strong>get things done.</strong></p><p><strong>🙌 If You Found This Helpful…</strong></p><p>If you enjoyed this breakdown of <strong>Agentic AI</strong> and learned something new:</p><p>👉 <strong>Give this article a clap (or many!)</strong> — it helps others discover it<br> 👉 <strong>Share it</strong> with your friends, colleagues, or anyone curious about AI<br> 👉 <strong>Follow me on Medium</strong> for more deep‑dives into AI, agents, reasoning systems, and the future of autonomous intelligence</p><p>Your support keeps me motivated to publish more high‑quality content.<br> <strong>Stay tuned — the next article in this Agentic AI series is coming soon! 🚀</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c53121adb71a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Banks Still Use Java (Even in 2026)]]></title>
            <link>https://medium.com/@pvprasanth474/why-banks-still-use-java-even-in-2026-bf5cf528f5ad?source=rss-f882d8a1979f------2</link>
            <guid isPermaLink="false">https://medium.com/p/bf5cf528f5ad</guid>
            <category><![CDATA[software]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[java]]></category>
            <category><![CDATA[banks]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[PV Prasanth]]></dc:creator>
            <pubDate>Fri, 16 Jan 2026 06:29:29 GMT</pubDate>
            <atom:updated>2026-01-16T06:29:29.125Z</atom:updated>
            <content:encoded><![CDATA[<p>And why they don’t jump to new “trendy” languages</p><figure><img alt="Why Banks Still Use Java" src="https://cdn-images-1.medium.com/max/930/1*7n-hVuHHQYT0Psj7B6nXMw.png" /></figure><p>When you hear about new programming languages like Python, Go, or Rust, it’s easy to think big companies — especially banks — would switch to them. But the reality is: <strong>banks still depend heavily on Java</strong>, a language created more than 30 years ago. And there are very good reasons for that.</p><p>Let’s break it down in a simple way.</p><h3>Java Is Everywhere in Big Companies</h3><p>Java is not just “old.” It’s one of the most trusted and widely used languages in the world.</p><ul><li>Over <strong>90% of Fortune 500 companies</strong> use Java.</li><li>Java holds <strong>25% of the global programming market</strong> and is used by <strong>402,000+ companies</strong>.</li><li>In banking, Java is the <strong>#1 language</strong> powering core banking transactions, ATM systems, and customer service apps.</li></ul><p>So, yes — banks <em>are</em> still choosing Java today.</p><h3>The Real Reason Banks Stick to Java</h3><p>Banks move <strong>trillions of dollars every day</strong>. If their system goes down for even a minute, the damage could be huge.</p><p>That’s why banks don’t choose languages because they’re “popular” or “trendy.” They choose what’s <strong>proven, stable, safe, and reliable</strong>.</p><h4>1. Java Can Handle Massive Workloads</h4><p>Bank systems process:</p><ul><li>Millions of transactions</li><li>Billions of records</li><li>24/7 real‑time operations</li></ul><p>Java’s JVM has been tested for decades in exactly these situations, and it performs very reliably.</p><p>Python, for example, struggles with heavy concurrency, and Go — while fast — doesn’t yet have decades of banking tools and libraries.</p><h4>2. Banks Need Stability and Security</h4><p>Banks deal with sensitive customer data and are watched closely by regulators.</p><p>They need systems that are:</p><ul><li>Easy to audit</li><li>Safe from security risks</li><li>Stable for 10–20+ years</li></ul><p>Java is designed with strong security features:</p><ul><li>Sandboxed execution</li><li>No dangerous pointers</li><li>Strong type safety</li><li>Automatic memory management</li></ul><p>These features make it a safe choice for financial systems</p><h4>3. Easy to Hire Java Developers</h4><p>Banks need thousands of developers across the world.</p><p>Java is taught in colleges everywhere, so companies can easily hire talent and build large teams.<br> Go, on the other hand, has a smaller developer pool — making it risky for long‑term banking projects.</p><h3>4. Java Has a Mature, Battle‑Tested Ecosystem</h3><p>Java has:</p><ul><li>Spring</li><li>Hibernate</li><li>Jakarta EE</li><li>Strong logging, monitoring, and messaging frameworks</li></ul><p>These tools were built and improved over decades, often inside banking environments. No new language offers this depth yet.</p><h3>Why Banks Don’t Switch to Python or Go</h3><p>To keep it simple:</p><p><strong>Python</strong></p><p>Great for:</p><ul><li>Data science</li><li>AI</li><li>Prototyping</li></ul><p>But not good for:</p><ul><li>Heavy transaction systems</li><li>High‑speed real‑time processing</li></ul><p>Because of the <strong>GIL</strong> and slower runtime</p><p><strong>Go</strong></p><p>Great for:</p><ul><li>Modern cloud systems</li><li>Fast, concurrent microservices</li></ul><p>But:</p><ul><li>Not enough experienced developers yet</li><li>Not many banking‑grade libraries or frameworks</li></ul><p>So banks use Go in some new fintech areas, but not for the core systems where stability is everything.</p><h3>Conclusion</h3><p>Banks don’t pick languages just because they’re new or cool.<br> They pick what keeps <strong>money moving safely and reliably</strong>.</p><p>And right now, <strong>Java is still the best tool for that job</strong> — proven, secure, stable, and supported by millions of developers.</p><p>That’s why, even in 2026, Java continues to be the backbone of the banking and payments world.</p><p>🙏 Thanks for Reading!</p><p>What do you think — should banks experiment more, or is “proven over trendy” still the right way to build financial systems? Let me know in the comments!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bf5cf528f5ad" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Is a Language Model, and Why Should You Care?]]></title>
            <link>https://medium.com/@pvprasanth474/what-is-a-language-model-and-why-should-you-care-4b4ce0c2930c?source=rss-f882d8a1979f------2</link>
            <guid isPermaLink="false">https://medium.com/p/4b4ce0c2930c</guid>
            <category><![CDATA[large-language-models]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[PV Prasanth]]></dc:creator>
            <pubDate>Thu, 19 Jun 2025 09:32:27 GMT</pubDate>
            <atom:updated>2025-06-19T09:32:27.804Z</atom:updated>
            <content:encoded><![CDATA[<p>We’re living in a golden age of artificial intelligence, where AI not only plays chess but also writes code, composes poetry, and explains black holes. At the heart of this revolution are <strong>Large Language Models (LLMs)</strong> — a term you’ve likely heard, but might not fully understand.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Tq4pE7S8LL_snYYldgPIuA.png" /><figcaption>Large language model(LLM’s)</figcaption></figure><p>In this post, I’ll demystify LLMs by breaking them down into three foundational concepts:</p><ul><li>What is a <strong>language model</strong>?</li><li>How does it relate to the broader field of <strong>Natural Language Processing (NLP)</strong>?</li><li>What makes a model “<strong>large</strong>”?</li></ul><h4>What is a Language Model?</h4><p>A <strong>Language Model (LM)</strong> is a AI system that predicts the next word in a sentence. That may sound simple — but when scaled up to billions of examples and trained with trillions of parameters, it becomes an incredibly powerful tool.</p><p>Think of it as <strong>autocomplete on steroids</strong>. If you say:</p><blockquote>“The capital of France is…”</blockquote><p>The model predicts “Paris” based on statistical patterns learned during training.</p><p>The model doesn’t “know” the world the way humans do — it learns patterns by seeing billions of examples.</p><h4>What is NLP, and Where Do Language Models Fit?</h4><p><strong>Natural Language Processing (NLP)</strong> is the field of AI that deals with understanding and generating human language. It covers a wide range of tasks:</p><ul><li><strong>Sentiment Analysis</strong> (is this review positive or negative?)</li><li><strong>Named Entity Recognition</strong> (spot the names of people or places)</li><li>Text Summarization</li><li>Machine Translation</li><li>Chatbots &amp; Question Answering</li></ul><p>Language models like GPT-4 are <strong>general-purpose engines</strong> that can perform all these tasks through prompting — no task-specific training needed.</p><h4>What Makes a Language Model “Large”?</h4><p><strong>Large Language Models (LLMs)</strong> are simply <strong>very big</strong> language models — trained on <strong>massive datasets</strong> using <strong>billions of parameters</strong> — allowing them to write, code, summarize, translate, and reason with astonishing fluency.</p><p>You’ve probably heard terms like “GPT-3 has 175 billion parameters.” But what does that mean?</p><p>Let’s break it down:</p><ul><li><strong>Parameters</strong>: These are the internal weights that the model adjusts during training. Think of them as the knobs and dials the model uses to “learn” how language works.</li><li><strong>Training data</strong>: LLMs are trained on <strong>massive</strong> datasets — books, websites, articles, code, conversations.</li><li><strong>Compute power</strong>: Training a modern LLM takes thousands of GPUs running for weeks or months.</li></ul><p>The “large” in LLM refers to:</p><ul><li>Size of the model (parameters)</li><li>Breadth of knowledge (data)</li><li>Depth of linguistic understanding</li></ul><blockquote>A language model doesn’t “think.” It predicts. But with enough training, those predictions start to feel a lot like intelligence. That’s the magic — and mystery — of modern AI.</blockquote><h4>Why It Matters?</h4><p>Understanding the basics of LLMs helps you use them better. You’ll learn:</p><ul><li>How to <strong>prompt</strong> effectively</li><li>When to trust the output — and when not to</li><li>What tasks they’re good at</li><li>Why they sometimes get things hilariously wrong</li></ul><p>Whether you’re using ChatGPT, Bard, Claude, or any other LLM, knowing how they work empowers you to use them as tools — not just toys.</p><h3>Conclusion:</h3><p>Large Language Models are the engines behind today’s most powerful AI applications. They’re statistical predictors at heart, yet capable of writing essays, solving math problems, and imitating human creativity.</p><p>🙏 Thanks for Reading!</p><p>Thanks so much for taking the time to read this post. I hope it helped demystify what language models and NLP really are, and why “large” models like GPT-4 are such a big deal.</p><p>f you found this helpful, feel free to <strong>👏 clap, comment, or share</strong> with your team or followers.— I’d love to hear your thoughts, questions, or experiences using LLMs.</p><p>Stay curious, and keep learning.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4b4ce0c2930c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[1 = 1 Really Necessary in SQL Queries]]></title>
            <link>https://medium.com/@pvprasanth474/1-1-really-necessary-in-sql-queries-5d9df5ab69e4?source=rss-f882d8a1979f------2</link>
            <guid isPermaLink="false">https://medium.com/p/5d9df5ab69e4</guid>
            <category><![CDATA[sql]]></category>
            <category><![CDATA[sql-injection]]></category>
            <dc:creator><![CDATA[PV Prasanth]]></dc:creator>
            <pubDate>Thu, 15 May 2025 07:18:55 GMT</pubDate>
            <atom:updated>2025-05-15T07:18:55.934Z</atom:updated>
            <content:encoded><![CDATA[<p>If you’ve been working with SQL or reading dynamic query generation code, you might have come across something that looks a bit odd:</p><pre>WHERE 1 = 1</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DhqJOH5Awzmeugb73wp0JQ.png" /><figcaption>1 = 1 in SQL?</figcaption></figure><p>At first glance, it seems pointless. After all, 1 <strong>does</strong> equal 1—always. So what’s the deal with this?</p><p>In this post, we’ll explore why 1 = 1 is a surprisingly useful SQL trick, especially when used in dynamic query generation. We&#39;ll look at some real-world examples go over the pros and cons, and answer the big question:</p><blockquote>Do you need 1 = 1 in your SQL queries—or is it just a quirky habit?</blockquote><h4>What Does 1 = 1 Mean in SQL?</h4><p>In SQL, 1 = 1 is a <strong>boolean expression that always evaluates to true — </strong>you could think of it as saying, “include everything.”</p><p>So this query:</p><pre>SELECT * FROM users WHERE 1 = 1;</pre><p>…returns all users, since 1 = 1 is always true.</p><p>Sounds pointless?</p><p>Not quite.</p><h4>Where It Shines: Building Dynamic SQL Queries</h4><p>The true power of 1 = 1 shows up when you&#39;re <strong>generating queries dynamically</strong>, like in backend code where filters are added based on user input.</p><p>Let’s say you’re building a product search with optional filters like:</p><ul><li>Category</li><li>Maximum price</li><li>Rating</li></ul><p>You don’t know ahead of time which filters the user will apply. This is where 1 = 1 helps streamline your logic.</p><h4>Example: Dynamic SQL with Optional Filters</h4><p><strong>With 1=1:</strong></p><pre>public String buildProductQuery(String category, Double maxPrice) {<br>    StringBuilder query = new StringBuilder(&quot;SELECT * FROM products WHERE 1=1&quot;);<br><br>    if (category != null &amp;&amp; !category.isEmpty()) {<br>        query.append(&quot; AND category = &#39;&quot;).append(category).append(&quot;&#39;&quot;);<br>    }<br><br>    if (maxPrice != null) {<br>        query.append(&quot; AND price &lt;= &quot;).append(maxPrice);<br>    }<br><br>    return query.toString();<br>}</pre><p>If the user selects:</p><ul><li>Category: &quot;Books&quot;</li><li>Max price: 30</li></ul><p>You end up with:</p><pre>SELECT * FROM products WHERE 1=1 AND category = &#39;Books&#39; AND price &lt;= 30</pre><p>The use of WHERE 1=1 makes it easy to <strong>append all filters with </strong><strong>AND</strong>, without worrying about whether a filter is the first one or not.</p><p><strong>Without 1 = 1:</strong></p><p>Let’s say you didn’t use 1 = 1. Now your code needs to know whether it’s appending the <strong>first filter</strong> (so it uses WHERE) or a <strong>subsequent one</strong> (so it uses AND). It ends up messier:</p><pre>StringBuilder query = new StringBuilder(&quot;SELECT * FROM products&quot;);<br>boolean hasCondition = false;<br><br>if (category != null) {<br>    query.append(&quot; WHERE category = &#39;&quot;).append(category).append(&quot;&#39;&quot;);<br>    hasCondition = true;<br>}<br><br>if (maxPrice != null) {<br>    if (hasCondition) {<br>        query.append(&quot; AND&quot;);<br>    } else {<br>        query.append(&quot; WHERE&quot;);<br>    }<br>    query.append(&quot; price &lt;= &quot;).append(maxPrice);<br>}</pre><p>This version is more complex and error-prone. Using 1 = 1 removes the need for tracking the state of the query—you just keep appending filters with AND.</p><h4>SQL Injection Risks</h4><p>Let’s be clear — 1 = 1 is <strong>not</strong> dangerous by itself.</p><p>But <strong>how</strong> you build queries <strong>can be dangerous</strong>. If you’re manually concatenating user input into SQL strings, you’re risking <strong>SQL injection</strong> attacks.</p><p>Vulnerable:</p><pre>String sql = &quot;SELECT * FROM users WHERE username = &#39;&quot; + username + &quot;&#39; AND password = &#39;&quot; + password + &quot;&#39;&quot;;</pre><p>If the user inputs:</p><ul><li>username = ‘admin’</li><li>password = ‘anything’ OR ‘1’ = ‘1’</li></ul><p>The resulting SQL becomes:</p><pre>SELECT * FROM users WHERE username = &#39;admin&#39; AND password = &#39;anything&#39; OR &#39;1&#39; = &#39;1&#39;</pre><p>Because &#39;1&#39; = &#39;1&#39; is always true, the attacker bypasses authentication.</p><p>Safe with Prepared Statements:</p><pre>String sql = &quot;SELECT * FROM users WHERE username = ? AND password = ?&quot;;<br>PreparedStatement stmt = conn.prepareStatement(sql);<br>stmt.setString(1, username);<br>stmt.setString(2, password);</pre><p>When using frameworks like JPA, Hibernate, or query builders, they often handle this safely for you.</p><h4>Do I Need 1 = 1 in Regular Queries?</h4><p>Nope! If you’re writing static queries directly in SQL, there’s <strong>no reason</strong> to include 1 = 1.</p><p>Instead of this:</p><pre>SELECT * FROM orders WHERE 1 = 1 AND status = &#39;shipped&#39;;</pre><p>Just write:</p><pre>SELECT * FROM orders WHERE status = &#39;shipped&#39;;</pre><p>1 = 1 is only helpful when <strong>conditions are added dynamically</strong>, not when writing queries by hand.</p><h4>When to Use 1 = 1</h4><p>Use it when:</p><ul><li>You’re <strong>building SQL queries dynamically</strong> (in application code).</li><li>You want to simplify the logic of adding multiple AND conditions.</li><li>You’re working with code templates where conditional logic is injected.</li></ul><p>Avoid it when:</p><ul><li>You’re writing static SQL directly in the database or a script.</li><li>You’re using a query builder or ORM that handles dynamic filtering for you.</li><li>You don’t need conditional logic in your SQL.</li></ul><h4>Conclusion</h4><p>At first glance, 1 = 1 in SQL seems like a useless trick. But it turns out to be a practical, time-saving technique—especially when building queries dynamically in backend code.</p><p>It lets you cleanly stack AND conditions without worrying about whether you&#39;re adding the first filter or not. That’s why it’s commonly used in languages like Java, Python, and PHP when building SQL strings based on user input.</p><p>It’s <strong>not just a quirky habit</strong> — it’s a practical trick for a specific use case. But <strong>you shouldn’t use it everywhere</strong>.</p><p>🙏 Thanks for Reading!</p><p>I hope this cleared up the mystery behind 1 = 1 in SQL and helped you understand <strong>when to use it—and when to skip it</strong>. If you found this helpful, feel free to <strong>👏 clap, comment, or share</strong> with your team or followers.</p><p>Have questions or tricks of your own? Drop them in the comments — I’d love to hear how <em>you</em> handle dynamic SQL!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5d9df5ab69e4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Quartz Scheduler in Java : Persistence with JDBCJobStore Explained]]></title>
            <link>https://medium.com/@pvprasanth474/quartz-scheduler-in-java-persistence-with-jdbcjobstore-explained-884572434e19?source=rss-f882d8a1979f------2</link>
            <guid isPermaLink="false">https://medium.com/p/884572434e19</guid>
            <category><![CDATA[quartz]]></category>
            <category><![CDATA[java]]></category>
            <category><![CDATA[spring-boot]]></category>
            <dc:creator><![CDATA[PV Prasanth]]></dc:creator>
            <pubDate>Mon, 12 May 2025 13:50:59 GMT</pubDate>
            <atom:updated>2025-05-13T06:07:28.809Z</atom:updated>
            <content:encoded><![CDATA[<h3>Quartz Scheduler in Java : Persistence with JDBCJobStore Explained</h3><p>Welcome back to the Quartz Scheduler series! In p<a href="https://medium.com/@pvprasanth474/core-concepts-of-quartz-50792dde8eb7">revious blog,</a> we explored core concepts: Jobs, Triggers, Schedulers, and Cron expressions. Now, it’s time to talk about <strong>persistence</strong> — a crucial part of running reliable, long-lived jobs in production.</p><blockquote><strong>Persistence</strong> — how Quartz stores your jobs so they can survive app restarts, failures, or even scale across servers.</blockquote><p>If you’re serious about job reliability or need clustering, <strong>JDBCJobStore</strong> is the tool you need.</p><h4>What Is Persistence in Quartz?</h4><p>Quartz provides two types of job stores:</p><p><strong>RAMJobStore</strong> (default): Keeps jobs in memory. Super fast, but everything disappears when your app shuts down.</p><p><strong>JDBCJobStore</strong>: Stores jobs in a relational database — durable, reliable, and restart-proof.</p><p>With JDBCJobStore, your job metadata lives in database tables. This means your scheduled jobs persist across deployments, crashes, and server reboots.</p><h4>Quartz Tables: What’s Stored Where?</h4><p>Once you configure Quartz for persistence, it creates and uses a set of standard tables. Here are the most important ones:</p><p>QRTZ_JOB_DETAILS — Stores job class, name, group, etc.</p><p>QRTZ_TRIGGERS — Links job to its trigger (schedule)</p><p>QRTZ_CRON_TRIGGERS — Stores cron expressions</p><p>QRTZ_SIMPLE_TRIGGERS — For simple intervals like “every 5 min”</p><p>QRTZ_FIRED_TRIGGERS — Tracks currently running jobs</p><p>QRTZ_LOCKS — Prevents duplicate firing in clustered environments</p><p>Optional tables like QRTZ_CALENDARS, QRTZ_JOB_LISTENERS, etc., are used for advanced scheduling needs.</p><h4>Table Lifecycle: What Happens and When?</h4><p>Here’s how these tables work at runtime:</p><p><strong>When a job is scheduled</strong>:</p><ul><li>Inserts into QRTZ_JOB_DETAILS, QRTZ_TRIGGERS, and either QRTZ_CRON_TRIGGERS or QRTZ_SIMPLE_TRIGGERS.</li></ul><p><strong>When a trigger fires</strong>:</p><ul><li>Entry is added to QRTZ_FIRED_TRIGGERS.</li><li>The job executes via a worker thread.</li></ul><p><strong>On success/failure</strong>:</p><ul><li>QRTZ_TRIGGERS is updated.</li><li>Misfire or retry logic (if any) is applied.</li></ul><p><strong>When jobs/triggers are deleted</strong>:</p><ul><li>Their rows are removed from all related tables.</li></ul><h4>Setting Up JDBCJobStore</h4><ol><li>Add Dependencies</li></ol><pre>&lt;dependency&gt;<br>  &lt;groupId&gt;org.quartz-scheduler&lt;/groupId&gt;<br>  &lt;artifactId&gt;quartz&lt;/artifactId&gt;<br>  &lt;version&gt;2.3.2&lt;/version&gt;<br>&lt;/dependency&gt;<br>&lt;dependency&gt;<br>  &lt;groupId&gt;org.postgresql&lt;/groupId&gt;<br>  &lt;artifactId&gt;postgresql&lt;/artifactId&gt;<br>  &lt;version&gt;42.5.1&lt;/version&gt;<br>&lt;/dependency&gt;</pre><p>2. Create Tables</p><p>Use the official <a href="https://github.com/quartz-scheduler/quartz/tree/main/quartz/src/main/resources/org/quartz/impl/jdbcjobstore">Quartz SQL scripts</a> to generate tables. Choose your DB (e.g., PostgreSQL or MySQL) and run the script.</p><p>3. quartz.properties Configuration</p><pre>org.quartz.scheduler.instanceName = MyQuartzScheduler<br>org.quartz.scheduler.instanceId = AUTO<br><br>org.quartz.jobStore.class = org.quartz.impl.jdbcjobstore.JobStoreTX<br>org.quartz.jobStore.driverDelegateClass = org.quartz.impl.jdbcjobstore.StdJDBCDelegate<br>org.quartz.jobStore.tablePrefix = QRTZ_<br>org.quartz.jobStore.dataSource = myDS<br><br>org.quartz.dataSource.myDS.driver = org.postgresql.Driver<br>org.quartz.dataSource.myDS.URL = jdbc:postgresql://localhost:5432/mydb<br>org.quartz.dataSource.myDS.user = user<br>org.quartz.dataSource.myDS.password = pass</pre><p>4. Start Scheduler</p><pre>SchedulerFactory factory = new StdSchedulerFactory(&quot;quartz.properties&quot;);<br>Scheduler scheduler = factory.getScheduler();<br>scheduler.start();</pre><p><strong>Conclusion</strong></p><p>JDBCJobStore brings durability and reliability to your scheduling layer. If your app needs to remember jobs across restarts or scale scheduling across a cluster, persistence is non-negotiable.</p><p>With Quartz + JDBC, you get full visibility, fine control, and high confidence — all while keeping job execution consistent.</p><p><strong>Thanks for reading! </strong>If this helped you, give it a 👏 or drop a comment. I’d love to hear how you’re using Quartz in your applications.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=884572434e19" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Core Concepts of Quartz]]></title>
            <link>https://medium.com/@pvprasanth474/core-concepts-of-quartz-50792dde8eb7?source=rss-f882d8a1979f------2</link>
            <guid isPermaLink="false">https://medium.com/p/50792dde8eb7</guid>
            <category><![CDATA[spring-boot]]></category>
            <category><![CDATA[java]]></category>
            <category><![CDATA[quartz]]></category>
            <dc:creator><![CDATA[PV Prasanth]]></dc:creator>
            <pubDate>Mon, 12 May 2025 12:34:27 GMT</pubDate>
            <atom:updated>2025-05-12T12:41:09.667Z</atom:updated>
            <content:encoded><![CDATA[<p>Welcome to the second part of our Quartz Scheduler series! In the <a href="https://medium.com/@pvprasanth474/introduction-to-quartz-scheduler-in-java-7de47bbaa991">previous </a>post, we covered Quartz fundamentals and its architecture. Now, let’s break down the <strong>four pillars</strong> of Quartz: <strong>Job</strong>, <strong>Trigger</strong>, <strong>Scheduler</strong>, and <strong>Cron Expressions</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/752/1*JSlSZQWBfwVBjSxwm7yIyg.png" /><figcaption>Core concepts of quartz</figcaption></figure><p>If you’re a Java developer looking to schedule background tasks with precision and flexibility, understanding these components is a must.</p><h4>1. Job — The Task to Be Executed</h4><p>What Is a Job?</p><p>A <strong>Job</strong> is a class that defines the work you want to run— like sending an email, generating a report, or cleaning up temp files. It must implement the org.quartz.Job interface.</p><p>Example: A Simple Job</p><pre>import org.quartz.Job;<br>import org.quartz.JobExecutionContext;<br>import org.quartz.JobExecutionException;<br><br>public class PrintDateJob implements Job {<br>    @Override<br>    public void execute(JobExecutionContext context) throws JobExecutionException {<br>        System.out.println(&quot;Current Time: &quot; + new java.util.Date());<br>    }<br>}</pre><p>When Quartz runs this job, it prints the current time.</p><h4>2. Trigger — The Scheduler’s Clock</h4><p>A <strong>Trigger</strong> decides <strong>when</strong> a job should run. You link it to a JobDetail.Quartz offers multiple trigger types:</p><ul><li>SimpleTrigger: For intervals like every 10 seconds.</li><li>CronTrigger: For calendar-like schedules (e.g., every Monday at 8 AM).</li></ul><pre>Trigger trigger = TriggerBuilder.newTrigger()<br>    .withIdentity(&quot;dailyTrigger&quot;, &quot;group1&quot;)<br>    .withSchedule(CronScheduleBuilder.cronSchedule(&quot;0 0 5 * * ?&quot;))<br>    .build();</pre><p><strong>Cron expression:</strong> 0 0 5 * * ? means: Run at 5:00 AM every day.</p><h4>3. Scheduler — The Orchestrator</h4><p>What Is a Scheduler?</p><p>The <strong>Scheduler</strong> is the core of Quartz. It accepts jobs and triggers, then manages their execution over time.</p><pre>// Create Scheduler<br>SchedulerFactory factory = new StdSchedulerFactory();<br>Scheduler scheduler = factory.getScheduler();<br>scheduler.start();<br><br>// Schedule the job<br>scheduler.scheduleJob(jobDetail, trigger);</pre><p>This starts the scheduler and registers the job with its trigger.</p><h4>4. Cron Expressions — Fine-Grained Time Control</h4><p>Quartz supports powerful <strong>Cron expressions</strong> to run jobs with complex timing patterns.</p><pre>Seconds Minutes Hours Day-of-Month Month Day-of-Week Year(optional)</pre><p>Examples:</p><p>0 0 5 * * ? — Every day at 5:00 AM</p><p>0 0/10 * * * ? — Every 10 minutes</p><p>0 0 0 ? * MON — Every Monday at midnight</p><h4>Complete Working Example</h4><p>Let’s put it all together to print the current date every 10 minutes.</p><pre>import org.quartz.*;<br>import org.quartz.impl.StdSchedulerFactory;<br><br>public class QuartzExample {<br>    public static void main(String[] args) throws SchedulerException {<br>        JobDetail job = JobBuilder.newJob(PrintDateJob.class)<br>                .withIdentity(&quot;printDateJob&quot;, &quot;group1&quot;)<br>                .build();<br><br>        Trigger trigger = TriggerBuilder.newTrigger()<br>                .withIdentity(&quot;tenMinTrigger&quot;, &quot;group1&quot;)<br>                .withSchedule(CronScheduleBuilder.cronSchedule(&quot;0 0/10 * * * ?&quot;))<br>                .build();<br><br>        Scheduler scheduler = StdSchedulerFactory.getDefaultScheduler();<br>        scheduler.start();<br>        scheduler.scheduleJob(job, trigger);<br>    }<br>}</pre><h4>Real Use Cases:</h4><p>Report Generation — Schedule daily or weekly reports</p><p>Email Reminders — end alerts to users based on time</p><p>Batch Processing — Clean up logs or cache at night</p><h4>Tips &amp; Best Practices:</h4><ul><li>Don’t forget to shut down the scheduler gracefully.</li><li>Don’t put heavy logic directly inside execute() — delegate to services.</li><li>Name your jobs and triggers clearly.— useful for debugging and management.</li></ul><pre>scheduler.shutdown(true); // Waits for running jobs to finish</pre><h4>Conclusion</h4><p>By mastering <strong>Jobs</strong>, <strong>Triggers</strong>, <strong>Schedulers</strong>, and <strong>Cron expressions</strong>, you now have the essential building blocks to use Quartz confidently in Java.</p><p>It’s time to move beyond @Scheduled and take control of your application’s scheduling.</p><p>👉 In the next blog, we’ll cover <strong>Quartz Persistence (JDBCJobStore)</strong> — what it is, how to configure it, and how jobs survive restarts.</p><p><strong>🙏 Thanks for Reading!</strong></p><p>If you found this helpful, give it a 👏 or share it with your Java dev friends.</p><p>Have questions? Drop them in the comments — I’d love to hear what you’re building with Quartz.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=50792dde8eb7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Introduction to Quartz Scheduler in Java]]></title>
            <link>https://medium.com/@pvprasanth474/introduction-to-quartz-scheduler-in-java-7de47bbaa991?source=rss-f882d8a1979f------2</link>
            <guid isPermaLink="false">https://medium.com/p/7de47bbaa991</guid>
            <category><![CDATA[java]]></category>
            <category><![CDATA[quartz-scheduler]]></category>
            <category><![CDATA[spring-boot]]></category>
            <dc:creator><![CDATA[PV Prasanth]]></dc:creator>
            <pubDate>Mon, 12 May 2025 08:20:31 GMT</pubDate>
            <atom:updated>2025-05-12T08:20:31.425Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Why Do We Need a Scheduler?</strong></p><p>Imagine you run a system where you:</p><ul><li>Generate daily sales reports at midnight</li><li>Send out email reminders every 15 minutes</li><li>Sync data with external APIs every hour</li></ul><p>These are <strong>automated, recurring tasks</strong> — also known as <strong>jobs</strong>.</p><p>Java has some native ways like Timer, ScheduledExecutorService, or Spring’s @Scheduled — but maintaining, scaling, and persisting such jobs is messy.</p><p>That’s where <strong>Quartz Scheduler</strong> comes in.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GPgQCXgFOL48rDAyf_QJOQ.png" /><figcaption>Quartz</figcaption></figure><h4>What is Quartz?</h4><p><strong>Quartz</strong> is a richly featured, open-source job scheduling library in Java. It lets you:</p><ul><li>Run <strong>once</strong>, <strong>at intervals</strong>, or on <strong>complex cron expressions</strong></li><li>Are <strong>persistent</strong> across app restarts</li><li>Can be managed and updated <strong>dynamically at runtime</strong></li></ul><p><strong>Key Concepts:</strong></p><p><strong>Job </strong>— A class that contains your business logic</p><p><strong>Trigger</strong>— Defines <strong>when</strong> the job should run</p><p><strong>Scheduler </strong>— Coordinates jobs and triggers</p><p><strong>JobDetail</strong>— Holds metadata about the Job</p><p><strong>JobDataMap </strong>— Pass data to jobs</p><p><strong>Cron Expressions </strong>— Define complex schedules (like “every day at 5 AM”)</p><h4>Example — Print Date Every 10 Seconds:</h4><p>We’ll now build a simple Quartz job that prints the current date/time every 10 seconds</p><p>Step 1: Add Quartz Dependency</p><p>If you’re using Maven:</p><pre>&lt;dependency&gt;<br>  &lt;groupId&gt;org.quartz-scheduler&lt;/groupId&gt;<br>  &lt;artifactId&gt;quartz&lt;/artifactId&gt;<br>  &lt;version&gt;2.3.2&lt;/version&gt;<br>&lt;/dependency&gt;</pre><p>Step 2: Create the Job Class</p><p>This class contains the actual task we want to run.</p><pre>import org.quartz.Job;<br>import org.quartz.JobExecutionContext;<br>import org.quartz.JobExecutionException;<br>import java.time.LocalDateTime;<br><br>public class PrintTimeJob implements Job {<br>    @Override<br>    public void execute(JobExecutionContext context) throws JobExecutionException {<br>        System.out.println(&quot;Current Time: &quot; + LocalDateTime.now());<br>    }<br>}</pre><p>Step 3: Create the Scheduler with a Trigger</p><pre>import org.quartz.*;<br>import org.quartz.impl.StdSchedulerFactory;<br><br>import static org.quartz.JobBuilder.newJob;<br>import static org.quartz.TriggerBuilder.newTrigger;<br>import static org.quartz.SimpleScheduleBuilder.simpleSchedule;<br><br>public class QuartzExample {<br>    public static void main(String[] args) throws SchedulerException {<br><br>        // Step 1: Define the Job<br>        JobDetail job = newJob(PrintTimeJob.class)<br>                .withIdentity(&quot;timeJob&quot;, &quot;group1&quot;)<br>                .build();<br><br>        // Step 2: Define the Trigger<br>        Trigger trigger = newTrigger()<br>                .withIdentity(&quot;timeTrigger&quot;, &quot;group1&quot;)<br>                .startNow()<br>                .withSchedule(simpleSchedule()<br>                        .withIntervalInSeconds(10)<br>                        .repeatForever())<br>                .build();<br><br>        // Step 3: Schedule the job<br>        Scheduler scheduler = StdSchedulerFactory.getDefaultScheduler();<br>        scheduler.start();<br>        scheduler.scheduleJob(job, trigger);<br>    }<br>}</pre><p>Output:</p><p>Every 10 seconds, you’ll see:</p><pre>Current Time: 2025-05-12T13:25:30.423<br>Current Time: 2025-05-12T13:25:40.429<br>Current Time: 2025-05-12T13:25:50.431<br>...</pre><h4>Pros of Using Quartz:</h4><ul><li>Supports cron and complex schedules</li><li>Persistent job store with DB support</li><li>Scales across clusters</li><li>Rich lifecycle API</li><li>Dynamic job management</li><li>Exception handling and retries</li></ul><h4>When to Use Quartz:</h4><p>Use Quartz when you need:</p><ul><li>Complex or recurring schedules (e.g., Cron-based)</li><li>Persistent jobs across server restarts</li><li>Dynamic job updates at runtime</li></ul><h4>When <em>Not</em> to Use Quartz:</h4><p>Avoid Quartz if:</p><ul><li>You only need a few simple recurring tasks — use @Scheduled instead</li><li>Your job logic is lightweight and doesn’t need persistence</li><li>You prefer a zero-config or annotation-based approach</li></ul><h4>Real-World Analogy:</h4><p>Think of Quartz as the <strong>train station master</strong>:</p><ul><li><strong>Jobs</strong> are trains</li><li><strong>Triggers</strong> are the timetables</li><li><strong>Scheduler</strong> ensures trains run on time</li></ul><h4>Conclusion</h4><p>Quartz gives Java developers <strong>precise control over job scheduling</strong> — from simple timers to complex enterprise workflows.</p><p>In this post, we covered:</p><ul><li>What Quartz is and why it matters</li><li>Key components like Job, Trigger, Scheduler</li><li>A real Java example that prints time every 10 seconds</li><li>Pros, cons, and when to use it</li></ul><p>In the next part, we’ll explore <strong>Jobs, Triggers, Cron expressions, and how to pass data to jobs</strong> — with real use cases and implementation patterns.</p><h3>🙏 Thanks for Reading!</h3><p>If you found this post helpful, feel free to:</p><ul><li>💬 Leave a comment with your thoughts or questions</li><li>🔁 Share it with your developer friends</li></ul><p>Stay tuned for the next post — and happy scheduling! ⏱️🚀</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7de47bbaa991" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>