<?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 Kennedy Nwachukwu on Medium]]></title>
        <description><![CDATA[Stories by Kennedy Nwachukwu on Medium]]></description>
        <link>https://medium.com/@kenedynwachukwu?source=rss-792957a97e94------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*wC1_bAy2F8iWalhzzuNgdA.jpeg</url>
            <title>Stories by Kennedy Nwachukwu on Medium</title>
            <link>https://medium.com/@kenedynwachukwu?source=rss-792957a97e94------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 18:01:29 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@kenedynwachukwu/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[From Deployment to Security: Auditing My Azure Book Review Application with AI Assistance]]></title>
            <link>https://medium.com/@kenedynwachukwu/from-deployment-to-security-auditing-my-azure-book-review-application-with-ai-assistance-96ac72745b5c?source=rss-792957a97e94------2</link>
            <guid isPermaLink="false">https://medium.com/p/96ac72745b5c</guid>
            <category><![CDATA[idm]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[azure]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[claude-code]]></category>
            <dc:creator><![CDATA[Kennedy Nwachukwu]]></dc:creator>
            <pubDate>Thu, 03 Sep 2026 04:11:52 GMT</pubDate>
            <atom:updated>2026-09-03T11:59:11.115Z</atom:updated>
            <content:encoded><![CDATA[<h3>Your Application Is Deployed. But Is Your Infrastructure Secure?</h3><p>Deploying an application to the cloud is an exciting milestone. Seeing the frontend load, the backend communicate successfully, and the database store application data gives a sense of accomplishment.</p><p>But deployment is not the end of the process.</p><p>A more important question follows:</p><p><strong>Is the infrastructure supporting the application secure?</strong></p><p>That was the question I explored in <strong>Assignment 7 of my Week 7 Azure learning journey</strong>.</p><p>As part of my Week 7 project, I built and deployed a <strong>Book Review application on Microsoft Azure</strong> using a three-tier architecture. After deploying the application and its supporting infrastructure, I carried out a security posture audit to identify common cloud security misconfigurations.</p><p>This project gave me the opportunity to connect several important areas of Cloud Engineering and DevOps:</p><ul><li>Cloud infrastructure deployment</li><li>Network security</li><li>Bash scripting</li><li>Azure CLI</li><li>Security auditing</li><li>Agentic AI</li><li>Human-in-the-loop remediation</li></ul><h3>The Azure Infrastructure Behind My Book Review Application</h3><p>My Book Review application was deployed using a three-tier architecture.</p><p>The infrastructure consisted of separate layers to support different responsibilities within the application.</p><h3>Web Tier</h3><p>The web tier consisted of an Azure Virtual Machine responsible for serving the frontend application through <strong>Nginx</strong>.</p><p>This layer handles incoming web traffic and provides users with access to the application.</p><h3>Application Tier</h3><p>The application tier consisted of a separate Azure Virtual Machine running the backend application.</p><p>Separating the backend from the web tier allowed the application logic to operate independently from the frontend-serving layer.</p><h3>Database Tier</h3><p>The database layer used <strong>Azure Database for MySQL Flexible Server</strong>.</p><p>This provided a managed database service for storing and managing the application’s data.</p><h3>Network and Supporting Infrastructure</h3><p>The infrastructure also included:</p><ul><li>Network Security Groups (NSGs)</li><li>A Load Balancer</li><li>Azure Storage</li><li>Azure Virtual Machines</li><li>Managed Disks</li></ul><p>Network Security Groups were particularly important because they controlled which network traffic was allowed to reach different parts of the infrastructure.</p><p>After successfully deploying the application, I wanted to move beyond simply asking whether the infrastructure was working.</p><p>I wanted to ask:</p><blockquote><strong><em>Are the security configurations supporting this application correctly configured?</em></strong></blockquote><h3>Building a Read-Only Azure Security Audit</h3><p>I built a <strong>read-only Bash script</strong> to audit the Azure resources supporting my Book Review application.</p><p>The important design principle behind the script was simple:</p><blockquote><strong><em>The audit should collect evidence, not modify infrastructure.</em></strong></blockquote><p>Using the Azure CLI (az), the script inspected the Azure environment and generated a security posture report.</p><p>The audit focused on four key security checks.</p><h3>1. Network Security Group Rules</h3><p>The first check looked for unrestricted inbound access to management ports such as:</p><ul><li>SSH — Port 22</li><li>RDP — Port 3389</li></ul><p>Leaving management ports open to unrestricted internet access can expose infrastructure to automated scanning and login attempts.</p><p>The audit checked whether inbound rules allowed unrestricted access from sources such as:</p><ul><li>*</li><li>Internet</li><li>0.0.0.0/0</li></ul><p>A healthy configuration should restrict management access to known and authorized IP addresses or networks.</p><h3>2. Azure Storage Account Public Blob Access</h3><p>The second check verified whether public blob access was enabled on the Azure Storage Account.</p><p>Publicly accessible storage can potentially expose sensitive data if containers or blobs are accidentally configured for anonymous access.</p><p>The audit verified that public blob access was disabled.</p><h3>3. Managed Disk Encryption at Rest</h3><p>The third check examined the encryption configuration of the Virtual Machine’s managed disk.</p><p>Encryption at rest helps protect data stored on the underlying disk infrastructure.</p><p>The audit confirmed that the operating system disk was protected using Azure’s platform-managed encryption.</p><h3>4. Azure Database for MySQL Public Network Access</h3><p>The final check examined whether public network access was enabled for the Azure Database for MySQL Flexible Server.</p><p>Databases contain valuable application data, so limiting unnecessary public exposure is an important security consideration.</p><p>The audit verified the database network configuration and checked whether public access was disabled.</p><h3>Connecting the Audit to AI Assistance</h3><p>Building the Bash script was only part of the assignment.</p><p>The next step was to connect the audit workflow to <strong>Claude Code</strong> as a reusable /azure-audit skill.</p><p>This introduced an important AI-assisted workflow.</p><p>The Bash script was responsible for:</p><blockquote><strong><em>Gathering security evidence.</em></strong></blockquote><p>Claude Code was responsible for:</p><blockquote><strong><em>Analyzing the evidence and explaining the findings.</em></strong></blockquote><p>The AI-assisted skill could:</p><ul><li>Read the generated audit report</li><li>Identify PASS, WARN, and FAIL findings</li><li>Explain the security implications of a finding</li><li>Recommend a remediation command</li><li>Suggest a verification step</li></ul><p>However, there was an important restriction:</p><blockquote><strong><em>The AI was not allowed to automatically modify the Azure infrastructure.</em></strong></blockquote><p>This was the most important lesson I took from the assignment.</p><p>AI assistance does not have to mean giving an AI unrestricted permission to change production or cloud infrastructure.</p><p>Instead, AI can assist engineers with analysis and recommendations while the human remains responsible for approving and performing infrastructure changes.</p><h3>Testing the Audit with a Controlled Misconfiguration</h3><p>An audit script is only useful if it can successfully detect a problem.</p><p>To test the complete workflow, I temporarily introduced a controlled security misconfiguration into my Azure environment.</p><p>I created an unrestricted inbound SSH rule.</p><p>The rule allowed SSH access on port 22 from an unrestricted source.</p><p>When I ran the audit again, the result changed from a healthy posture to:</p><pre>FAIL: 1<br>Overall Status: FAIL</pre><p>The audit successfully identified the temporary unrestricted SSH rule.</p><p>This demonstrated that the script was not simply checking for healthy configurations — it could also detect the security condition it was designed to identify.</p><h3>AI Analysis Without AI Remediation</h3><p>After the audit detected the finding, I used the /azure-audit skill.</p><p>Claude Code analyzed the evidence and explained why unrestricted SSH access could be a security concern.</p><p>The analysis highlighted potential risks such as:</p><ul><li>Brute-force login attempts</li><li>Credential-stuffing attacks</li><li>Automated internet scanning</li><li>Unauthorized access attempts</li></ul><p>The AI then recommended a remediation command.</p><p>However, it did not execute the command.</p><p>This distinction is important.</p><p>The workflow was designed so that:</p><p><strong>The AI analyzes.<br> The human decides.<br> The human performs the change.<br> The system verifies the result.</strong></p><p>After reviewing the recommendation, I manually removed the temporary unrestricted SSH rule myself.</p><h3>Verifying the Fix</h3><p>Security remediation should not end immediately after making a change.</p><p>The final step is verification.</p><p>After manually removing the temporary rule, I ran the audit again.</p><p>The final result was</p><p>✅<strong>PASS: 4</strong><br> ⚠️ <strong>WARN: 0</strong><br> ❌ <strong>FAIL: 0</strong><br> ✅ <strong>Overall Status: HEALTHY</strong></p><p>The second audit provided evidence that the temporary security finding had been resolved.</p><p>I also preserved reports showing the different stages of the process:</p><ul><li>Baseline report</li><li>Before-fix report showing the security finding</li><li>After-fix report showing the healthy result</li></ul><p>This made it possible to demonstrate the complete lifecycle of the finding.</p><h3>The Security Workflow</h3><p>The most important concept behind this project was the workflow:</p><h3>Gather → Analyze → Human Fix → Verify</h3><h3>1. Gather</h3><p>Collect security evidence using read-only commands.</p><p>The audit script examines the Azure environment without changing resources.</p><h3>2. Analyze</h3><p>Use AI assistance to interpret the evidence.</p><p>The AI explains findings, identifies risks, and recommends possible remediation.</p><h3>3. Human Fix</h3><p>The human operator reviews the recommendation and performs the infrastructure change.</p><p>This ensures that critical cloud resources are not modified automatically without human oversight.</p><h3>4. Verify</h3><p>Run the audit again.</p><p>The final verification confirms whether the remediation successfully resolved the original finding.</p><h3>What I Learned About AI in Cloud and DevOps</h3><p>One of the biggest lessons from this project was that AI assistance in Cloud Engineering and DevOps does not necessarily mean automation without control.</p><p>AI can provide value by helping engineers:</p><ul><li>Analyze large amounts of information</li><li>Interpret security findings</li><li>Explain potential risks</li><li>Recommend remediation steps</li><li>Improve operational efficiency</li></ul><p>But critical infrastructure decisions can still remain under human control.</p><p>This creates a more controlled model of automation and AI assistance:</p><blockquote><strong><em>AI provides intelligence and recommendations. Humans maintain responsibility for actions.</em></strong></blockquote><p>For cloud security and infrastructure operations, this approach can be especially valuable because infrastructure changes can have significant consequences.</p><h3>Bringing the Week 7 Project Together</h3><p>This assignment connected the different stages of my Azure project.</p><p>I started by building infrastructure to support a Book Review application.</p><p>I deployed:</p><ul><li>Web infrastructure</li><li>Application infrastructure</li><li>Database infrastructure</li><li>Network security controls</li><li>Supporting Azure resources</li></ul><p>I then moved beyond deployment and evaluated the security posture of the environment.</p><p>The complete journey looked like this:</p><p><strong>Build Infrastructure → Deploy Application → Audit Security → Detect Misconfigurations → AI-Assisted Analysis → Human Remediation → Verify the Fix</strong></p><p>This project gave me practical experience in understanding that Cloud and DevOps work does not stop when an application becomes available.</p><p>Deployment is only one stage.</p><p>Engineers also need to think about:</p><ul><li>Security</li><li>Monitoring</li><li>Configuration</li><li>Access control</li><li>Continuous auditing</li><li>Verification after changes</li></ul><h3>Conclusion</h3><p>My Week 7 Azure Assignment 7 was a valuable hands-on experience because it combined cloud deployment with security auditing and AI assistance.</p><p>I was able to audit the same Azure infrastructure I built to support my Book Review application and test a complete security workflow from detection through remediation and verification.</p><p>The project reinforced three important lessons for me:</p><p><strong>First:</strong> Cloud infrastructure should be audited, not simply deployed and forgotten.</p><p><strong>Second:</strong> AI can provide valuable assistance in analyzing security evidence and recommending actions.</p><p><strong>Third:</strong> Human oversight remains important when making changes to cloud infrastructure.</p><p>The Azure CLI commands and technologies may change from one cloud environment to another, but the underlying discipline remains valuable:</p><blockquote><strong><em>Gather evidence. Analyze findings. Keep humans in control of remediation. Verify the result.</em></strong></blockquote><p>This project strengthened my practical experience with <strong>Microsoft Azure, Cloud Security, Bash scripting, Azure CLI, Agentic AI, and Human-in-the-Loop automation</strong> as I continue building my skills in Cloud Engineering and DevOps.</p><p>P.S. This post is part of the DevOps Micro Internship with Agentic AI Cohort 3 by Pravin Mishra. You can begin your DevOps journey by joining the DMI waiting list. (<a href="https://forms.gle/3hvrWJBDzsDeJoPs6">https://forms.gle/3hvrWJBDzsDeJoPs6</a>)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ycH4Tf7VOqZ33cpiJZYZ0g.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YpD1hCnHfxclubvwlO3vcg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*a14onE2_K4TcvR2y2Nngaw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*I3O9BXfDHoZ5ZSBUwoOMbg.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=96ac72745b5c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Reflection — Week 2]]></title>
            <link>https://medium.com/@kenedynwachukwu/reflection-week-2-a3471bb4b9e4?source=rss-792957a97e94------2</link>
            <guid isPermaLink="false">https://medium.com/p/a3471bb4b9e4</guid>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[devops-workflow]]></category>
            <category><![CDATA[terraform]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <dc:creator><![CDATA[Kennedy Nwachukwu]]></dc:creator>
            <pubDate>Mon, 13 Jul 2026 16:20:40 GMT</pubDate>
            <atom:updated>2026-09-03T11:48:03.789Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>By: Kennedy Nwachukwu</strong></p><p>If I had to describe Week 2 in one sentence, I would say it completely changed the way I see AI in software engineering.</p><p>Going into this week, I thought AI was simply something that could generate code or answer technical questions. I never imagined it could be configured to work like a real teammate with responsibilities, memory, rules, and even restrictions. That was probably the biggest mindset shift for me.</p><p>The first thing that stood out was <strong>Claude Code</strong>. I spent quite a bit of time learning how it works beyond just chatting with it. Creating the CLAUDE.md file made me realize how important context is. Instead of repeating instructions every session, I can give Claude a permanent understanding of my project. It felt similar to onboarding a new developer to a team—you don&#39;t just tell them what to build; you explain how the project is structured and the standards they should follow.</p><p>Another topic I found interesting was <strong>Skills and Subagents</strong>. At first, I didn’t fully understand why one AI would need different agents. But after looking at agents like the security auditor, cost optimizer, and Terraform writer, it started making sense. Each one has a specific responsibility, just like different engineers on a DevOps team. That separation of duties makes the workflow more organized and more reliable.</p><p>The area that challenged me the most was <strong>Hooks and Permissions</strong>. Honestly, I spent hours troubleshooting JSON errors, Python command issues, and hook configurations. There were moments when I thought I had everything set up correctly, only to discover that a missing comma or a small syntax mistake was stopping the entire configuration from working. On top of that, I ran into GitHub push errors because Terraform provider files were too large, and later my Claude API credits ran out in the middle of the assignment. None of those issues were part of the “happy path,” but they taught me an important lesson: in DevOps, troubleshooting is just as valuable as building.</p><p>I also learned about <strong>Memory</strong>, which I found surprisingly useful. Seeing Claude remember project decisions across sessions made me understand why documentation is so important. Instead of relying on memory — or explaining the same things repeatedly — you can create a system that preserves important decisions for future work.</p><p>One habit I want to build from this week is documenting important decisions immediately instead of assuming I’ll remember them later. Whether it’s updating CLAUDE.md, keeping project memory clean, or writing meaningful Git commit messages, I want future me—and anyone else working on the project—to understand why decisions were made.</p><p>Looking back, this wasn’t the easiest week. I spent more time debugging than I expected, and there were moments when things became frustrating. But I can honestly say those struggles taught me more than if everything had worked perfectly the first time. They forced me to slow down, read error messages carefully, research solutions, and think like an engineer instead of someone looking for a quick fix.</p><p>Week 2 has shown me that becoming a DevOps Engineer isn’t just about learning tools. It’s about developing the discipline to build systems that are reliable, secure, and maintainable. More importantly, it’s about learning how to work with AI intentionally rather than depending on it blindly. That is a lesson I’ll carry with me throughout the rest of this internship and beyond.</p><p>P.S. This post is part of the DevOps Micro Internship with Agentic AI Cohort 3 by Pravin Mishra. You can begin your DevOps journey by joining the DMI waiting list. <a href="http://P.S. This post is part of the DevOps Micro Internship with Agentic AI Cohort 3 by Pravin Mishra. You can begin your DevOps journey by joining the DMI waiting list. (https://forms.gle/3hvrWJBDzsDeJoPs6)">(https://forms.gle/3hvrWJBDzsDeJoPs6)</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LkNms56E5eojdYciFoRecQ.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a3471bb4b9e4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Version 2.0: The Engineer Who Chose to Build]]></title>
            <link>https://medium.com/@kenedynwachukwu/version-2-0-the-engineer-who-chose-to-build-96c119783709?source=rss-792957a97e94------2</link>
            <guid isPermaLink="false">https://medium.com/p/96c119783709</guid>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[career-growth]]></category>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <dc:creator><![CDATA[Kennedy Nwachukwu]]></dc:creator>
            <pubDate>Mon, 29 Jun 2026 15:46:50 GMT</pubDate>
            <atom:updated>2026-09-03T12:00:13.272Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9051l9_lNS9MAIy5r4kHAw.png" /></figure><p><em>“Three years ago, Kennedy Nwachukwu was just another ambitious learner trying to find his place in tech. Today, he is known as a DevOps Engineer who builds reliable cloud systems, mentors aspiring engineers, and creates solutions that help businesses grow.”</em></p><p>People often see the results but rarely the journey behind them.</p><p>Before the certifications, the job offers, and the successful projects, there were countless hours of learning, failed deployments, debugging sessions, and moments of doubt. Instead of giving up, Kennedy treated every challenge as another lesson. Every setback became an opportunity to improve.</p><p>Rather than collecting certificates, he focused on building.</p><p>He designed cloud infrastructure on AWS, automated deployments with Jenkins and GitHub Actions, deployed applications using Docker and Kubernetes, and managed infrastructure with Terraform. His GitHub portfolio became a timeline of his growth, filled with real-world DevOps projects that demonstrated practical problem-solving instead of theoretical knowledge.</p><p>His commitment to learning earned him industry-recognized certifications and opened doors to work on engineering teams where he improved deployment speed, strengthened cloud security, automated repetitive tasks, and increased system reliability.</p><p>But what made his journey truly meaningful wasn’t the technology — it was the people.</p><p>Kennedy shared his knowledge through technical articles, open-source contributions, webinars, and mentorship. He helped aspiring engineers build confidence, create strong portfolios, and launch careers in cloud computing and DevOps.</p><p>His passion for solving problems also led him to co-found technology startups that simplified infrastructure management and software delivery, helping organizations scale with greater efficiency.</p><p>When people asked him what made the difference, his answer was always simple:</p><p><strong>“Consistency beats intensity. Build more than you consume, solve real problems, and never stop learning.”</strong></p><p>Looking back, Kennedy realized that becoming a DevOps Engineer was only part of the journey. The real achievement was becoming someone who created value through the systems he built, the people he mentored, and the problems he solved.</p><p>This wasn’t just a career transformation.</p><p>It was proof that small, consistent actions, repeated over time, can completely change the direction of a person’s life.</p><p>P.S. This post is part of the DevOps Micro Internship with Agentic AI Cohort 3 by Pravin Mishra. You can begin your DevOps journey by joining the DMI waiting list. (<a href="https://forms.gle/3hvrWJBDzsDeJoPs6">https://forms.gle/3hvrWJBDzsDeJoPs6</a>)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=96c119783709" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>