<?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 Oshin Mundada on Medium]]></title>
        <description><![CDATA[Stories by Oshin Mundada on Medium]]></description>
        <link>https://medium.com/@oshinmundada?source=rss-2aa2546acaf2------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*aNWUsBs5fN1Op8yH.</url>
            <title>Stories by Oshin Mundada on Medium</title>
            <link>https://medium.com/@oshinmundada?source=rss-2aa2546acaf2------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 20:16:00 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@oshinmundada/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[If movies were to be made on programmer’s life…]]></title>
            <link>https://medium.com/@oshinmundada/if-movies-were-to-be-made-on-programmers-life-2b62ec0b3753?source=rss-2aa2546acaf2------2</link>
            <guid isPermaLink="false">https://medium.com/p/2b62ec0b3753</guid>
            <category><![CDATA[programmer]]></category>
            <category><![CDATA[movies]]></category>
            <category><![CDATA[humor]]></category>
            <category><![CDATA[programmers-life]]></category>
            <category><![CDATA[geek]]></category>
            <dc:creator><![CDATA[Oshin Mundada]]></dc:creator>
            <pubDate>Sat, 05 May 2018 00:41:02 GMT</pubDate>
            <atom:updated>2018-05-05T00:41:02.998Z</atom:updated>
            <content:encoded><![CDATA[<p>(Titles inspired from famous movies)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*45f-v_Y5CtP9shORXWr35w.jpeg" /><figcaption>The Matrix (cos some movies are named just about right)</figcaption></figure><h4>1.​ The Perks of Being a Semicolon-Free Language | The Perks of Being a Wallflower</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/1*2NKRituUL06QlfhJdIHwZg.png" /></figure><h4>2. Owl of the Tech Street | Wolf of the Wall Street</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/450/1*sAwd0UbqrUSOWYsH76RHpQ.jpeg" /></figure><h4>3. Lonelihood of the Coding Pants | Sisterhood of the Traveling Pants</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/357/1*SpGyLmkUuFaV4KtCzHplRA.jpeg" /></figure><h4>4. Debug me if you can | Catch me if you can</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/625/1*9ppUPSlgrfCrygNiReDS-g.jpeg" /></figure><h4>5. The Sound of Mouse-Click | The Sound of Music</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*jR2MgiqBhsyDtE0emiQy5A.gif" /></figure><h4>6. Eternal Backlight of the Laptop Screen | Eternal Sunshine of the Spotless Mind</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/628/1*AToysVLjX91sEe9cusvslw.jpeg" /></figure><h4>7. Almost Compiled | Almost Famous</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/497/1*R7GRJeZrNQkPpaigX0Y9OA.jpeg" /></figure><h4>8. How to train your neural net | How to train your dragon</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/551/1*V3gcrzkm4Odc3xjV5X8dJQ.jpeg" /></figure><h4>9. The Fault in Our Code | The Fault in Our Stars</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/418/1*1AQi-60nEMJYUwOPAE24bA.jpeg" /></figure><h4>10. The Curious Case of Semicolon | The Curious Case of Benjamin Button</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/564/1*uN0205k_QuXY8GuOnTy2WA.jpeg" /></figure><h4>11. Requiem for a Solution | Requiem for a Dream</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/625/1*jCiqyLFLP6dMgyRkyVzT6w.jpeg" /></figure><h4>12. Mark Zuckerburg and the Facebook of Secrets | Harry Potter and the Chamber of Secrets</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/1*Z3OPN555S8I1DHNZU5p1WA.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2b62ec0b3753" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Experiences from my internship in machine learning]]></title>
            <link>https://medium.com/@oshinmundada/experiences-from-my-internship-in-machine-learning-b37f146c9dd8?source=rss-2aa2546acaf2------2</link>
            <guid isPermaLink="false">https://medium.com/p/b37f146c9dd8</guid>
            <category><![CDATA[internships]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[tensorflow]]></category>
            <category><![CDATA[intern]]></category>
            <dc:creator><![CDATA[Oshin Mundada]]></dc:creator>
            <pubDate>Mon, 30 Apr 2018 19:27:11 GMT</pubDate>
            <atom:updated>2018-04-30T20:38:23.033Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/250/1*bX_xOIf_bVRdlBJLqu5SVg.jpeg" /></figure><p>In the summer of 2017, I had the opportunity to intern at <a href="http://www.instrumentationlaboratory.com/us/en">Instrumentation Laboratory</a> on a machine learning project. For a short background, the company makes medical instruments. Other than their actual purpose, these instruments log out data based on some protocols, about the working of their different parts (like arm, barcode, fluidics, etc.). Our project was to indicate when there is a need for technical support in light of an improperly functioning instrument.</p><p>At the end of three months and my internship, I mused on some interesting feelings about the project. *Flipping a coin to decide whether to start with the good or the bad… and like always, bad triumphs 😈*</p><p>So here goes…</p><h4>1. We were chasing the algorithm, when the data hit us.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/583/1*WjWla4nGngNeUUdLhdbNRg.gif" /></figure><p>As the general conception, I always thought it was model fitting and testing that was the difficult and time consuming part. Turns out, the real deal is the data. We spent majority of our efforts in getting the data and pre-processing it — converting it to 0 and 1 matrices of date and weight segregated by area.</p><h4>2. Having lots of data is not enough. To apply machine learning techniques successfully, you need ground-truth data.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/498/1*ZSiW4QlQ3P4T1i-BDJwIlA.gif" /></figure><p>Machine learning is not magic. A model is only as good as you train it to be. A critical requirement for that was to have ‘clean’ data (data that indicated error-free working of device) and ‘anomalous’ data (data that indicated potential device failure). Getting hands on such data and labeling the training files was certainly not an easy task.</p><h4>3. First, you need to understand what you are working on.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/400/1*K8fhWj4waq67KXt2MnIT8g.gif" /></figure><p>The above gif is a very appropriate depiction of how quickly our excitement turned into a cloud of questions. Understanding the data, the working of the instruments, the protocols that were used to log the files and how and why they were faceted the way they were, played an important role in our decisions on how to model our network, initial weights and checking the accuracy of our training data. Questions are not so bad until they take 3 days or more to be answered. We depended on people from Boston, Barcelona and some ex-employees too. Between that and everyone’s availabilities, work was often hindered.</p><p>— — — — — — — — — — — — — — —</p><p>However, it wasn’t all bad. So let’s move on to the good:</p><h4>1. Tensorflow is a knight in shining armor.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/449/1*iYYsrLKZJT6icdPg0gu-gg.gif" /></figure><p>With a slightly steep learning curve, Tensorflow, Google’s open source framework for deep learning is super powerful. It catered to our modeling of an autoencoder neural network with 31 input and output nodes and 3 hidden layers. With substantial documentation, Tensorboard and efficient performance, the key is to play around till you learn the library’s power and then attack.</p><h4>2. Machine learning always has a lot (more) to offer.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/1*1CaXIx9fbVp58j0ZT3sh0w.gif" /></figure><p>With machine learning, we not only detected anomalies in the instrument logs, but also created a prediction model to foretell when the instrument will fail by utilizing sequence mining.</p><h4>3. The opportunity to work in this badass field.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*DwRWuan9cVEuKgoHmnZAow.gif" /></figure><p>It was the perfect job; you know the one that gets you excited for reasons more than being able to pay bills. I got to work in the leading-edge field of AI/ML, learn some pretty awesome things and finish with the joy of having conquered an aspired goal.</p><p>See more details of the internship <a href="https://docs.google.com/presentation/d/1cP43YbP0GtKw2jxuiQPz1ykK8WSL75fhm_7cOud6esg/edit#slide=id.g253ab468ae_0_0">here</a>.</p><h4>BONUS:</h4><p>A business trip to Boston! 😍</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/540/1*zKDJZhgE5CJv_Qg6uiwDrQ.jpeg" /></figure><p>Hope you enjoyed reading my experience. Every internship makes for a story and I would love to know yours. Share it in the comments below. :)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b37f146c9dd8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[5 things I learnt at my first job out of school]]></title>
            <link>https://medium.com/@oshinmundada/5-things-i-learnt-at-my-first-job-out-of-school-dbded331d20d?source=rss-2aa2546acaf2------2</link>
            <guid isPermaLink="false">https://medium.com/p/dbded331d20d</guid>
            <category><![CDATA[learning-to-code]]></category>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[best-practices]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[first-job]]></category>
            <dc:creator><![CDATA[Oshin Mundada]]></dc:creator>
            <pubDate>Sun, 29 Apr 2018 02:13:29 GMT</pubDate>
            <atom:updated>2018-04-29T02:13:29.477Z</atom:updated>
            <content:encoded><![CDATA[<p>Other than the fact, that there are no spring, summer and winter breaks at jobs, and people spend months counting down to their next vacation, I learnt some coding practices significant enough to share with all of you.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*hdqZt74SGKYoBWB_urWgqw.jpeg" /></figure><h4>1. Know your text editor for code</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/384/1*ey4A5EZYP5CkQP3rsz3MnQ.gif" /></figure><p>Code editors are powerful, make use of them. Learn the functionalities they provide. Also, there is not enough time to reach out for your mouse multiple times. Learn the keyboard shortcuts.</p><h4>2. Never forget to delete unused code</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/836/1*qNmZ-kcuiZPJmg4qUEnvZA.gif" /></figure><p>While code, in general, does not take a whole lot of space, it adds cost to your program. Chunks of commented code increases confusion and decreases readability, a combination you never want to have as a programmer. Forgetting to remove unused headers or imports is even worse, as it loads every time the component loads.</p><h4>3. Never keep the user loading</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*c9ZgwDNam0BmF-5qnEI5sw.gif" /></figure><p>Load pages with as light data as possible. Make the heavy fetch calls to server as and how the components on the page mount. Use spinners to let your users know the page is still working. If your code fails, make sure it fails gracefully.</p><h4>4. Force yourself to write reusable components</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/160/1*RyW-eR5EPq49igWl506lNg.gif" /></figure><p>Doing this has multiple advantages — helps keep the length of files shorter, makes code easier to read, leading to easier debuging and reduced development time, increased extensibility. <br><em>P.S. Do not underestimate the role of Rihanna in coding </em>🎧</p><h4>5. Do not get attached to your code</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/1*6pRWutJUajR4spIknibGew.gif" /></figure><p>There will be times when you worked on something for three hours and then came up with a totally new solution. In such cases, do not be afraid to hit the delete button. Never write ‘hacky’ code just to get it working with your old solution. There’s a high chance you might write the new code in much lesser time anyway. And if not immediately, you’ll surely save time in the future.</p><p>That’s all for my first post folks. Hope this quick read will help you. :)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dbded331d20d" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>