<?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 Vinny Joseph on Medium]]></title>
        <description><![CDATA[Stories by Vinny Joseph on Medium]]></description>
        <link>https://medium.com/@joseph.vinnyt?source=rss-f927b2651c03------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*fxtpaIGaV5bxuIv_tIdqYw.jpeg</url>
            <title>Stories by Vinny Joseph on Medium</title>
            <link>https://medium.com/@joseph.vinnyt?source=rss-f927b2651c03------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 00:42:27 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@joseph.vinnyt/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[The Spark — The Moment I Realized Anyone Can Build a Meaningful App]]></title>
            <link>https://medium.com/@joseph.vinnyt/the-spark-the-moment-i-realized-anyone-can-build-a-meaningful-app-2f49d48be7fd?source=rss-f927b2651c03------2</link>
            <guid isPermaLink="false">https://medium.com/p/2f49d48be7fd</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[faith]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Vinny Joseph]]></dc:creator>
            <pubDate>Tue, 28 Oct 2025 03:13:46 GMT</pubDate>
            <atom:updated>2025-10-28T15:19:19.849Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AGtFri1WizVjmHOWN7C4Jg.jpeg" /><figcaption>Photo by Viktoria: <a href="https://www.pexels.com/photo/fireworks-761547/">https://www.pexels.com/photo/fireworks-761547/</a></figcaption></figure><h3>The Spark — The Moment I Realized Anyone Can Build a Meaningful App</h3><p>It was 8 a.m. on September 17, 2025, and I was at the LA Fitness Gym — my usual routine after dropping my daughter off at school. Once I’m out, there are no excuses.</p><p>As always, I started my workout with the familiar playlist: the daily Bible blessing video my mom sends me from India, an NPR news update, a quick AI podcast, and then a TED talk to finish the elliptical session.</p><p>That morning’s talk was titled <strong>“</strong><a href="https://youtu.be/Cf6CHpLbf5w?si=_i5SgFJuDWztlX_-"><strong>How to Build a 100x Career with AI</strong></a><strong>”</strong> by <em>Ishan Sharma</em>. I’ve watched plenty of TED and TEDx talks before — most inspire me, but this one <em>stayed with me</em>.</p><h3>When Curiosity Turns Into a Calling</h3><p>The talk wasn’t complicated — it was real. Ishan spoke about how someone from rural India could connect with global innovators, simply by learning and building with AI. He mentioned an interview with Notion’s founder and how they look for people with a <em>product mindset</em>.</p><p>But what truly caught my attention was when he started listing tools that helped him build fast — tools that made creating an app or a startup more accessible than ever.</p><p>One of them was <a href="https://replit.com/refer/josephvinnyt"><strong>Replit</strong></a> — a platform I had heard about but never tried. When he showed how he built a complete website within hours, something clicked for me. Maybe the gap between “idea” and “product” wasn’t as wide as I thought.</p><p>As I left the gym, I felt a <em>renewed drive</em> — the excitement to build something new.</p><h3>The First Step</h3><p>When I finished my work later that evening, I decided to give it a try. I created a Replit account, not knowing what I was going to build.</p><p>Then I remembered my mom’s daily Bible verses — and a question she had asked me several times:</p><blockquote><em>“Can’t I get verses and reflections that connect to what I’m going through each day?”</em></blockquote><p>That question became my <strong><em>why</em></strong>.</p><p>Within minutes, I drafted a simple prompt and used AI tools to generate the first version of what would become <a href="https://qa.evangify.co/"><strong>Evangify</strong></a><strong> (</strong><em>QA Preview, but open for anyone curious to explore and share feedback.</em><strong>)</strong> — a spiritual companion app that offers personalized verses, reflections, and actions tailored to each person’s daily life and challenges.</p><h3>Main Tools That Made It Possible</h3><p>(A few of these are referral links — they help support my writing at no extra cost to you.)</p><p>I didn’t write thousands of lines of code or set up complicated servers. Instead, I built it using a combination of modern tools that anyone can access:</p><ul><li><a href="https://replit.com/refer/josephvinnyt"><strong>Replit</strong></a> — for coding, hosting, and versioning (its <em>Ghostwriter</em> feature was a huge help)</li><li><a href="https://manus.im/invitation/UXG1ZX8WAUYXY"><strong>Manus</strong></a><strong> </strong>— my AI strategy assistant for market research, feature testing, and marketing strategy exploration etc</li><li><a href="https://platform.openai.com/"><strong>OpenAI API</strong></a> — for generating reflections and daily spiritual insights</li><li><a href="https://bible-api.com/"><strong>Bible API</strong></a> (open source) — to fetch daily verses dynamically</li><li><a href="https://claude.ai/"><strong>Claude</strong></a> — to brainstorm content tone and structure</li><li><a href="https://chatgpt.com/"><strong>ChatGPT</strong></a> — to refine UX, debug, and guide design logic</li></ul><p>These tools felt like a creative orchestra — each doing a small part to bring the idea to life.</p><h3>Rediscovering My Creative Spark</h3><p>That evening, I couldn’t stop. I sat back at my desk — coding, prompting, experimenting. I even skipped dinner with my kids and continue coding.</p><p>But that feeling of creative flow — the joy of building something from scratch — was priceless. As someone who’s spent years moving from coding to engineering management, I had almost forgotten that thrill.</p><p>By midnight, I had a working version ready. I shared it with my mom and brother back in India. They loved it — and, of course, immediately started sending me new ideas to build next.</p><p>That’s when I started to feel: this wasn’t just a side project. It was some thing more — to use technology to bring people closer to faith, reflection, and compassion.</p><h3>What Came After</h3><p>Since that night, I’ve been steadily refining <strong>Evangify</strong> — improving its design, adding new features, and preparing it for beta launch. All this happens during quiet early mornings or late evenings and weekends outside of work hours — the times I use to recharge my creativity and learn new tools hands-on.</p><p>Along the way, I learned valuable lessons about:</p><ul><li>How to use AI not just for productivity, but as a creative partner in coding and problem-solving</li><li>How to navigate challenges with tools like Replit and integrate them efficiently</li><li>Why designing for <em>real people</em> matters more than perfecting theoretical “personas”</li><li>And yes — the surprisingly fun (and sometimes painful) process of preparing for App Store submission</li></ul><p>I’ll be sharing more of these lessons and progress updates here as the journey continues.</p><p>What’s been most rewarding is how much these experiences have shaped my professional growth. Experimenting with AI tools and rapid prototyping in my side project has helped me bring a more innovative mindset to my day job — from exploring automation ideas to mentoring my team on emerging technologies.</p><p>Building <a href="https://qa.evangify.co/"><strong>Evangify</strong></a> reminded me that personal curiosity and professional growth don’t have to be separate. The lessons I’ve learned in those off-hours have directly influenced how I lead, make decisions, and encourage creativity within my engineering team.</p><p>Each step reminded me that we live in an incredible time — one where you can turn an idea into a working prototype in a matter of hours.</p><h3>My Takeaway</h3><p>If there’s one message I’d leave you with, it’s this:</p><blockquote><em>If you have an idea that solves even a small problem in your life, build it.<br>You might be surprised how far you can go with today’s tools.</em></blockquote><p>That 17-minute TED talk by Ishan Sharma reignited something I had almost lost: <em>the joy of creating</em>. This experience revived not only my passion for building but also my appreciation for how personal exploration outside work can fuel innovation within it.</p><p>And for me, that was <strong>The Spark</strong>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2f49d48be7fd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Large Language Model — 101 (Beginners)]]></title>
            <link>https://medium.com/@joseph.vinnyt/large-language-model-101-beginners-a97fa6b054b1?source=rss-f927b2651c03------2</link>
            <guid isPermaLink="false">https://medium.com/p/a97fa6b054b1</guid>
            <category><![CDATA[neural-networks]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[natural-language-process]]></category>
            <category><![CDATA[large-language-models]]></category>
            <dc:creator><![CDATA[Vinny Joseph]]></dc:creator>
            <pubDate>Mon, 03 Feb 2025 04:32:26 GMT</pubDate>
            <atom:updated>2025-02-03T16:19:04.611Z</atom:updated>
            <content:encoded><![CDATA[<h3>Large Language Model — 101 (Beginners)</h3><p>Large Language Models (LLMs) are basically super-smart computer programs that are excellent with words. They’re a type of AI, but instead of focusing on images or numbers, LLMs are all about understanding and using language.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*x9f9OyY1EzR4e7Uod19IWQ.jpeg" /></figure><h3><strong>Real-World LLM Examples:</strong></h3><ul><li><strong>ChatGPT:</strong> You’ve probably chatted with something like it. These chatbots can hold conversations, answer questions, and even write different kinds of creative content. It’s like texting a really knowledgeable friend.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*CK7NYAb9m-XHM9NZeJXtHA.jpeg" /></figure><h3><strong>How Do LLMs Work? (In Simple Terms)</strong></h3><p>Imagine teaching a dog a new trick. You repeat the command, reward the right action, and eventually, the dog gets it. LLMs learn in a similar way, but instead of treats, they get <em>data</em>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Jzh6FmHucFHEzoUEkVcZGw.jpeg" /></figure><ol><li><strong>Training Data:</strong> This is like the dog’s training manual. It’s a massive collection of text — books, websites, articles — that the LLM reads to learn the rules of language. The more data, the smarter the LLM gets.</li><li><strong>Neural Networks:</strong> This is the LLM’s brain. It’s a complex network of connections that process the training data. Think of it like a web where each strand is connected to others. This network helps the LLM understand the relationships between words.</li><li><strong>Parameters:</strong> These are like the dog’s memory of what it learned. They store all the information the LLM has gathered from the training data. The more parameters, the better the LLM remembers and understands.</li><li><strong>Fine-Tuning:</strong> Sometimes you want your dog to learn a <em>specific</em> trick. Fine-tuning is like that. It’s tweaking the LLM’s parameters to make it really good at a particular task, like summarizing news articles or translating from Spanish to English.</li></ol><h3><strong>Key Concepts in LLM Architecture</strong></h3><ul><li><strong>Tokens:</strong> Imagine breaking a sentence down into individual words and punctuation marks. That’s kind of what “tokens” are. They’re the smallest pieces of language that LLMs work with. LLMs convert text into these tokens so they can understand and process it. Think of them as the fundamental building blocks of language for LLMs.</li><li><strong>The Transformer:</strong> Think of the Transformer as a special type of LLM architecture. It’s like a super-efficient engine that makes LLMs even better. A key part of the Transformer is the “attention mechanism.” This helps the LLM focus on the most important parts of the text, like a spotlight highlighting key words in a sentence. The Transformer architecture, with its attention mechanisms, has been a significant breakthrough in Natural Language Processing.</li></ul><h3><strong>What Can LLMs Do?</strong></h3><p>LLMs are incredibly versatile. Here are some of their talents:</p><ul><li><strong>Text Generation:</strong> They can write all sorts of things, from poems and stories to emails and articles.</li><li><strong>Text Summarization:</strong> Need the gist of a long article? LLMs can condense it down for you.</li><li><strong>Language Translation:</strong> They can translate between languages, breaking down communication barriers.</li><li><strong>Question Answering:</strong> Ask them a question, and they’ll try their best to give you a helpful answer.</li><li><strong>Code Generation:</strong> They can even write computer code!</li></ul><h3><strong>Where Are LLMs Used?</strong></h3><p>LLMs are popping up everywhere!</p><ul><li><strong>Content Creation:</strong> Helping writers, marketers, and even musicians.</li><li><strong>Customer Service:</strong> Powering chatbots that answer your questions 24/7.</li><li><strong>Education:</strong> Creating personalized learning experiences for students.</li><li><strong>Healthcare:</strong> Analyzing medical records and assisting with research.</li><li><strong>Business:</strong> Automating tasks and helping with decision-making.</li></ul><h3><strong>LLM Limitations (They’re Not Perfect)</strong></h3><p>While LLMs are impressive, they’re still under development.</p><ul><li><strong>Lack of True Understanding:</strong> They’re really good at recognizing patterns in text, but they don’t truly <em>understand</em> what they’re reading. It’s like a parrot that can repeat phrases without knowing what they mean.</li><li><strong>Bias:</strong> If the training data contains biases, the LLM might also be biased.</li><li><strong>Limited Memory:</strong> They can sometimes forget details from earlier in a conversation.</li><li><strong>Computational Costs:</strong> Training and running these models can be expensive.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/528/1*gdQjmB3lVpkLzkdcy1AGZQ.png" /></figure><h3><strong>Conclusion</strong></h3><p>LLMs represent a significant advancement in artificial intelligence, offering powerful tools for understanding and generating human language. From powering chatbots to assisting in research, their applications are vast and constantly expanding. While limitations like bias and a lack of true understanding remain, ongoing research and development are continually pushing the boundaries of what’s possible. As LLMs mature, their impact on how we communicate and interact with technology will only continue to grow.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a97fa6b054b1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Path to Google Cloud Professional Cloud Architect Certification]]></title>
            <link>https://medium.com/priceline-labs/path-to-google-cloud-professional-cloud-architect-certification-da58da63492c?source=rss-f927b2651c03------2</link>
            <guid isPermaLink="false">https://medium.com/p/da58da63492c</guid>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[cloud]]></category>
            <category><![CDATA[gcp]]></category>
            <category><![CDATA[gcp-certification]]></category>
            <category><![CDATA[architecture]]></category>
            <dc:creator><![CDATA[Vinny Joseph]]></dc:creator>
            <pubDate>Fri, 27 Aug 2021 16:59:17 GMT</pubDate>
            <atom:updated>2021-08-27T16:59:17.307Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bYeLsevymg_84hWe1POldg.png" /><figcaption>Photo by <a href="https://unsplash.com/@mitchel3uo?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Mitchell Luo</a> on <a href="https://unsplash.com/s/photos/google-cloud?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>The Beginning</h3><p>It all began when Priceline started the journey to GCP (Google Cloud Platform). I was a novice with a lot of advanced cloud concepts. I had some grasp around AWS, as that was the cloud platform most of my professional network colleagues worked on, but I really missed the focused specialization. Luckily, I got the opportunity to be involved in the initial discussions with 12 Factor specifications and to work with the organization leaders, which really helped me to advance my understanding.</p><p>However, as we got into the actual hands on implementation, things started to get tricky. There was a lot of ambiguity in the arena.</p><ul><li><strong>Lack of terminology to describe requirements:</strong> Since the entire company was starting the cloud journey at the same time, there was some confusion at the beginning to describe the requirements with the right cloud terms.</li><li><strong>Choosing a database for a cloud-native application: </strong>This became really tricky as GCP offered different solutions based on the type and usage of the data.</li><li><strong>Navigating GCP in the dark:</strong> When we request GCP resources like databases, cloud storage, etc., we have only a limited understanding of access control, networking, etc. As a developer, this provided some abstraction to focus on the task at hand, but I was still curious to find out how things were actually done.</li></ul><h3>The Decision</h3><p>As I started my research to clarify requirements and to better understand concepts, I kept on hitting online resources leading to GCP certifications.</p><p>To an extent, the goal of a GCP certification is to ensure we have the right understanding of how GCP and its resources operate in order to make better decisions about how to leverage these resources. I slowly came to realize that working on a certification would advance my understanding not only of GCP but also of cloud architecture more generally. Hence, around mid-October 2020, I made the decision to start preparing for the GCP <a href="https://cloud.google.com/certification/cloud-architect">Professional Cloud Architect exam</a>.</p><h3>The Preparation</h3><p>Just like any software engineer, I started by Googling resources and tips on how to prepare for the GCP exam. Interestingly enough, I used a couple of articles on <em>Medium</em> to start with. Below are some resources that helped in my six months of preparation, during which I spent at least an hour each day studying.</p><p><strong>Medium Articles</strong></p><ul><li><a href="https://medium.com/tag/gcp-certification">https://medium.com/tag/gcp-certification</a></li></ul><p><strong>GCP Flowcharts</strong></p><ul><li><a href="https://grumpygrace.dev/posts/gcp-flowcharts/">https://grumpygrace.dev/posts/gcp-flowcharts/</a></li></ul><p><strong>GCP in Four words or less</strong></p><ul><li><a href="https://github.com/gregsramblings/google-cloud-4-words">https://github.com/gregsramblings/google-cloud-4-words</a></li></ul><p><strong>YouTube Videos</strong></p><ul><li>Google Cloud Tech -<a href="https://www.youtube.com/user/googlecloudplatform">https://www.youtube.com/user/googlecloudplatform</a></li><li>Learn with Mahesh -<a href="https://www.youtube.com/channel/UCL8vwy2jhEtQrC-Rx6qzCZg">https://www.youtube.com/channel/UCL8vwy2jhEtQrC-Rx6qzCZg</a></li><li>Cloud Advocate — <a href="https://www.youtube.com/playlist?list=PLBRBRV08tHh1e9oqTsONC5AuhFd-SUXWq">https://www.youtube.com/playlist?list=PLBRBRV08tHh1e9oqTsONC5AuhFd-SUXWq</a></li><li>Cloud Minute Videos -<a href="https://www.youtube.com/playlist?list=PLIivdWyY5sqIij_cgINUHZDMnGjVx3rxi">https://www.youtube.com/playlist?list=PLIivdWyY5sqIij_cgINUHZDMnGjVx3rxi</a></li><li>GCP Conference videos — <a href="https://www.youtube.com/playlist?list=PLIivdWyY5sqIXvUGVrFuZibCUdKVzEoUw">https://www.youtube.com/playlist?list=PLIivdWyY5sqIXvUGVrFuZibCUdKVzEoUw</a></li><li>Cloud Guru Courses — <a href="https://acloudguru.com/course/google-certified-professional-cloud-architect">https://acloudguru.com/course/google-certified-professional-cloud-architect</a></li></ul><p><strong>Google Sample Questions</strong></p><ul><li><a href="https://cloud.google.com/certification/sample-questions/cloud-architect">https://cloud.google.com/certification/sample-questions/cloud-architect</a></li></ul><p>Once I gained a better understanding of the various topics, the next step was to sit for some practice exams, which one can easily find online. A simple Google search returns a ton of resources, both free and paid. Many questions don’t have any incorrect answers but instead, have a “most correct” solution. The goal of the practice tests is not to learn the answers, but to enhance our thinking to identify the actual problem and find the best solution to solve the problem based on the given constraints.</p><h3>Before the Online Exam</h3><p><strong>Registration</strong></p><p>The first step is to <a href="https://webassessor.com/googlecloud">register for the exam</a>. The site was a bit confusing at first, but I got used to it over time.</p><p>You can reschedule your test date without incurring any extra fees until 72 hrs before the exam. After this, a fee may apply. A good strategy is to schedule your exam ahead and, a week before or so, try the Google practice test. If confident with the results, continue with the date, if not, reschedule. The registration fee is $200.</p><p><strong>New Online Exam Format</strong></p><p>The Remote Exam was introduced during the pandemic, replacing the onsite exams. I booked a Sunday morning 9:00 AM slot, so I could use half of the weekend for preparation and still enjoy a weekend day after the exam.</p><p><strong>Installation of Exam Software</strong></p><p>This is an important step and time-consuming. You need to install a software called <em>Sentinel</em> in order to launch the exam. Once launched, this will take over the computer, and you won’t be able to access any other applications. It is better to close all the apps before the exam. It also asks to take a photograph, which is used for verification before the exam. I did this on the day before the exam and lost precious time I could have used for my exam preparation. Ensure to install the software early.</p><p><strong>Clear The Space</strong></p><p>GCP exams require the exam room and computer desk to be cleared. You can only have your water bottle and the Government ID. I found it easier to use a different room than my home office for the exam, as it was much easier to clear. Ensure you have the room set up ahead of the exam; nothing on the table, floor, or walls. It is better to have a mirror ready. I will explain why in the next section.</p><p><strong>The Scanning</strong></p><p>The exam becomes available 10 minutes prior to the scheduled time. The first step is to scan your face and then the room. Once the pre-screening steps are completed, an agent will be online and message you from a chat window. The agent will ask to enable the camera and microphone, if it is not on yet, and to start scanning the Government ID. The next step is to slowly move around the room to show all the walls, corners, ceilings, and floor. The agent may ask questions based on what they see, but nothing was asked in my case. The agent will message you with the next action. I read somewhere they might also speak directly instead of messaging, so don’t let this catch you off guard. Finally, you have to show the back of the laptop as well as the screen using a mirror. This is to verify you don’t have any notes on the screen or anywhere else on the laptop. You have to keep the camera and the microphone turned on for the entire duration of the exam.</p><p>The entire process took around 20 minutes. After this step, there was again a wait to actually start the exam. I am not sure if this is expected, but all these delays felt a bit overwhelming. From launching the exam to the actual start it took around 30 minutes. Luckily, once the exam started, I shifted my focus to the exam and almost forgot about the pre-verification process.</p><h3>The Exam</h3><p>The exam itself contains 50 questions and takes two hours to complete. Google does not share an official pass mark for the GCPCA exam, but various resources say the passing score is 80%. The results won’t reveal the exact score or any details about the performance.</p><p>The questions were similar in structure to the practice questions Google and other resources offered. Each question has 4 different answer choices. The question will tell how many right answers to choose. You can answer the questions right away or mark them for a later review. The answers can be changed before the final submission.</p><p>Read each question and, more importantly, the answer choices carefully. Many times the difference between choices are just a few words around resource capacity, regional vs. zonal, database types, access roles, etc.</p><p>The case studies are really important. I had around fifteen questions from case studies. At the time of my exam, the three case studies were Mountkirk games, Dress4Win, and Terram-Earth.</p><p>I went through all the 50 questions in 70–80 minutes and took the remaining time to review all of my answers. Finally, with almost 15 minutes to spare, I clicked Submit.</p><h3>The Result</h3><p>The official result will take up to seven days. Luckily, you will get a provisional result right after submission. And yes, I passed! As I mentioned earlier, test takers won’t receive any information regarding performance or the score, and the result shown immediately after the exam is provisional. There are some reviews Google does before the release of the official result. After two days, I got the official email from Google with the result and a link to retrieve the digital certificate. You also get credits to buy certain items from the Google Cloud Certification Perk store. I ended up getting a laptop bag.</p><h3>The Next Steps</h3><p>Looking back, my journey to professional cloud certification was fascinating. I was always interested in system design and architecture and am still a big fan of books like <em>Clean Architecture </em>by Uncle Bob, but the exam preparation provided a new perspective on how I approached system design and architecture. Previously I always focused on the best technical solution, but the exam made me rethink my approach and consider the requirements from the business, functional, and technical perspectives. The GCP case studies are good examples of this.</p><p>I also realized that there isn’t any universally right solution and proper software architecture is one that strikes a balance between the cost, performance, and timeline objectives.</p><p>Another revelation was the discovery of a lot of common problems and similar solutions/patterns that are available across different platforms. For example, the use cases where Google pub-sub helps can also be supported by AWS Kinesis or Azure Event Hub in those respective platforms. In a non-cloud environment, this can be supported by Apache Kafka or RabbitMQ. In fact, Google made clear in their exam objective that the goal is to gain knowledge on the cloud domain, not just on GCP.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/445/1*sbFr4AA83JD0-HddDkfvlg.jpeg" /></figure><h3>Conclusion</h3><p>For many, the certification will be the end of a quest, but for me, it opened up a new world of exploration and excitement. I feel this is just the beginning. With my improved knowledge and understanding, I am looking to approach software architecture and system design problems from a fresh perspective and apply the learnings to my day-to-day work. I am also looking forward to the new developments and trends in the cloud domain as I continue to advance my learnings. I hope, sharing my experience will inspire someone to start their certification journey!</p><p>Excited by what we’re doing at Priceline? Visit <a href="https://careers.priceline.com/search?q=Technology">Priceline’s career site</a> to learn more about our mission, culture, and open positions.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=da58da63492c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/priceline-labs/path-to-google-cloud-professional-cloud-architect-certification-da58da63492c">Path to Google Cloud Professional Cloud Architect Certification</a> was originally published in <a href="https://medium.com/priceline-labs">priceline labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My Journey to TDD]]></title>
            <link>https://medium.com/priceline-labs/my-journey-to-tdd-38bd18241e02?source=rss-f927b2651c03------2</link>
            <guid isPermaLink="false">https://medium.com/p/38bd18241e02</guid>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[test-driven-development]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Vinny Joseph]]></dc:creator>
            <pubDate>Thu, 19 Dec 2019 20:36:14 GMT</pubDate>
            <atom:updated>2019-12-19T20:36:14.755Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="A laptop with some code showing — we hope these are tests!" src="https://cdn-images-1.medium.com/max/1024/1*e1iypsouAGB6jABtPW-JyA.jpeg" /></figure><p>Despite having heard a lot about Test Driven Development (TDD) during my career, spanning over a decade, I had never practiced TDD until a few months ago. As a personal challenge, I decided to experiment with TDD as one of my quarterly goals. Hence, in addition to the usual weight loss and self-improvement goals, I set a new goal to do TDD this year.</p><p>Historically, I always found reasons to evade TDD.</p><h3>Ways I have avoided TDD</h3><h4>TDD is almost impossible on legacy code.</h4><p>It is really hard to do TDD on legacy code due to the old nature of the code base, the lack of modularization, heavy usage of global variables, hardwired dependencies and lack of existing unit tests.</p><h4><strong>TDD is time-consuming.</strong></h4><p>TDD requires thought and analysis prior to writing code, and is thus time-consuming.</p><h4><strong>TDD means changing habits.</strong></h4><p>It became a habit to write code by thinking while you type. It is always hard to change a habit or to start a new one.</p><h4><strong>The product owners won’t understand the value of TDD.</strong></h4><p>Are the product owners willing to compromise on the delivery time in order to add more unit tests?</p><h4><strong>Peer pressure.</strong></h4><p>What will my teammates think when I suddenly start to behave abnormally, i.e. doing TDD?</p><h4><strong>TDD is not practical.</strong></h4><p>How can I consider TDD practical as it adds time to deliver high priority tasks and delays features, which in turn slows revenue growth?</p><h4>How I approached TDD</h4><p>Although I had many reasons to feel discouraged, I decided to try TDD on my next task. I had to add a bunch of validation rules to some legacy code and recognized that it was a good opportunity to try my first-ever TDD, which went as follows:</p><ol><li>Using the single responsibility principle and abstraction, I isolated the new changes to their own class. This way the changes to the legacy class became minimal. I ended up having a new class with one simple boolean, no argument function, which always returns false.</li><li>I wrote a unit test for this new function, which failed as expected.</li><li>I began adding code to the new function, with the goal of making the test pass. As I added more code, I thought through how to make the code testable. This resulted in loosely coupled code with dependency injections.</li><li>Once my initial test passed, I added more unit tests to cover the additional test cases, adding code to fix bugs as they arose. And whenever I made code changes, I reran all my tests to ensure the changes didn’t cause side effects.</li><li>After my code was complete and all tests were passing, I started to refactor things. I reran the tests each time I made changes. Since I had great coverage on my code, I was able to refactor my code with confidence.</li></ol><p>The end result surprised me. I had written some of my cleanest, most readable code to date. The simplicity of the code and its test coverage also surprised my teammates. Since then, I have used TDD many times, consistently producing code that I have great confidence in.</p><h3><strong>Here are some tips to getting started with TDD:</strong></h3><h4><strong>Start Simple</strong></h4><p>Always start simple. TDD differs substantially from the traditional development process. You will stumble on a lot of questions as you start to think about the code you are about to write. This is normal and it is one of the benefits of TDD. It makes us think about the design and dependencies earlier in the process. However, don’t let this confuse you. Start simple; write one failing test and work on making it pass.</p><h4><strong>Try Top-Down</strong></h4><p>It is much easier to try a top-down approach when doing TDD as opposed to a bottom-up approach. This means starting with the very high-level details and dependencies that we already know. Don’t worry about the minutiae of the code or the layout at the beginning. Once we have the tests in place, the code can be refactored.</p><h4><strong>Think First</strong></h4><p>Think before you type! This forces you to think about the code before getting started. While this introduces a delay in initial development cycle, the results will save time in the long run.</p><h4><strong>Continuous Refactoring</strong></h4><p>Once you have the initial test passing, the next stage is refactoring. This is where we can think more about the design principles and patterns, then apply them. There is always some uncertainty involved when it comes to refactoring code. However, with full test coverage, this becomes much easier and less concerning. Tests will ideally catch any bugs caused by refactoring. Always run the tests after each change to identify side effects right away.</p><h4><strong>Treat Unit Tests as Documentation</strong></h4><p>Consider the unit tests to be documentation. Ensure proper naming and descriptions for the tests. Clearly specify what we are testing and what is the expectation.</p><h3><strong>Conclusion</strong></h3><p>Changing habits always presents a challenge. The traditional way of coding became so habitual to most of us. I’ve started doing a lot more TDD since I first began practicing it, but I still appreciate the difficulty of taking the first step. I continue to find myself tempted to start coding in the traditional way, thinking while I type. But after gaining some momentum, I found myself writing clean and elegant code with excellent test coverage. If you have yet to try TDD, give it a shot. It’s worth your time!</p><p>Photo by Christopher Gower on Unsplash</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=38bd18241e02" width="1" height="1" alt=""><hr><p><a href="https://medium.com/priceline-labs/my-journey-to-tdd-38bd18241e02">My Journey to TDD</a> was originally published in <a href="https://medium.com/priceline-labs">priceline labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>