<?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 Mohita Parashar on Medium]]></title>
        <description><![CDATA[Stories by Mohita Parashar on Medium]]></description>
        <link>https://medium.com/@mohitaparashar?source=rss-7b1d59d20040------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Z2e1Xi6UGLmx-VUFxp73Fg.png</url>
            <title>Stories by Mohita Parashar on Medium</title>
            <link>https://medium.com/@mohitaparashar?source=rss-7b1d59d20040------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 14:18:02 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mohitaparashar/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[Why Bing should be your first choice for Health related queries!]]></title>
            <link>https://mohitaparashar.medium.com/why-bing-should-be-your-first-choice-for-health-related-queries-f9ad40d33a33?source=rss-7b1d59d20040------2</link>
            <guid isPermaLink="false">https://medium.com/p/f9ad40d33a33</guid>
            <dc:creator><![CDATA[Mohita Parashar]]></dc:creator>
            <pubDate>Mon, 14 Feb 2022 12:53:17 GMT</pubDate>
            <atom:updated>2022-04-27T14:53:48.780Z</atom:updated>
            <content:encoded><![CDATA[<p>A lot of people ask me,<br>“<em>Hey, what do you do on Bing?”<br>“Are you guys still innovating in Bing?”<br>“Who uses Bing anyway?”<br></em>and the most legit doubt<em>, <br>“Do you guys google too, when you can’t figure something out?”</em></p><p>Hello everyone!<br>I am a software developer at Microsoft and I work in the Health segment on Microsoft Bing. A lot of my friends are taken aback when I tell them I work on Bing, and it is probably not very hard to guess why!</p><p>Google has been successful in establishing itself as a household verb, used in everyone’s daily life. We have been so habituated to googling everything, that we are unaware of all the goodness that exists elsewhere. And today, I am going to tell you about one such segment, where statistically, people have loved the Bing answers more than Google’s and where you should try Binging first - the health segment.</p><p>Here’s why Bing should be your first choice for all the health related queries -</p><ul><li>User friendly one stop answers — which will save your time and effort of browsing through various web pages.</li><li>Rich experiences and easy to understand language — The information is organized topic wise and presented in short bullet points, making it very readable and understandable.</li><li>Ask health professionals directly — you can get your health related questions answered directly by a verified health professional.</li><li>Breaking the language barrier — the answers are available in 65 different languages, making it easier for people all around the world to understand.</li><li>Trustworthiness and Authoritative data — All the data in our experiences are reviewed by a panel of doctors.</li></ul><p>Let me walk you through a few of our experiences to demonstrate what I’ve been talking about.</p><ol><li><strong>Health Conditions related queries</strong></li></ol><p>We provide <strong>one stop solution</strong> for all the possible information you need to know about a health condition. The experience is enriched with helpful images and some of the most common <strong>FAQs answered by verified health professionals</strong>.</p><p>Different tabs present in this answer lets you know the symptoms, causes, complications, diagnosis, specialist to consult, nutritional recommendations and possible everything you need to know about this medical condition.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oYN8N2pP_fABAYHv9FD5Dw.png" /></figure><p><strong>2. Medical Symptoms</strong></p><p>We recognize the various health symptoms ranging from headache and stomach ache to shortness of breath, and provide you information regarding the causes, treatment, doctors and the related medical conditions, at one stop.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3Ia8k8kT1Eh2YPGTrZrfEw.png" /></figure><p>3. <strong>Health Medicine related queries</strong></p><p>We are here for all your medicine related queries.<br>We provide carefully curated data about medicines, what they treat, their common brands, side effects, FAQs related to the medicine and a lot more.</p><p>Again here, the different tabs present will serve all the possible queries you might have about a medicine.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_22s7cNx40qCoFQIx8uI2Q.png" /></figure><p>4.<strong> Medical Tests and Surgeries</strong></p><p>We have wide range of answers curated for helping you answer all the queries you might have regarding a medical test or surgery. We help you take an informed decision describing the risks, requirements, expectations and a lot more information concerning medical tests and surgeries.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*skeXqlSkmHXr7BsFgyAEPw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*10FbmWt_8D4TDUjsTTn8ng.png" /></figure><p>So <em>“Who uses Bing anyway?”</em>— Currently, millions of unique users.</p><p>“<em>Hey, what do you do on Bing?”<br>“Are you guys still innovating in Bing?”</em> — We continue to impact millions of lives by providing the above mentioned kinds of credible answers. We will continue to provide a wide range of answers catering to your health and wellness.</p><p>and lastly,</p><p><em>“Do you guys google too, when you can’t figure something out?” — </em>Well, we love google too, but for health queries, it is always Bing! :-)</p><p>Now, I can’t wait for you all to go and check out our work in the health segment on Bing and let me know if you would prefer binging your health related queries first!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f9ad40d33a33" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The one where I converted “Multiple Rejections” to “Multiple Offers”]]></title>
            <link>https://mohitaparashar.medium.com/the-one-where-i-converted-multiple-rejections-to-multiple-offers-4f9c2804b5f7?source=rss-7b1d59d20040------2</link>
            <guid isPermaLink="false">https://medium.com/p/4f9c2804b5f7</guid>
            <category><![CDATA[job-interview]]></category>
            <category><![CDATA[microsoft]]></category>
            <category><![CDATA[amazon]]></category>
            <category><![CDATA[system-design-interview]]></category>
            <category><![CDATA[coding-interviews]]></category>
            <dc:creator><![CDATA[Mohita Parashar]]></dc:creator>
            <pubDate>Mon, 31 May 2021 17:22:51 GMT</pubDate>
            <atom:updated>2021-05-31T17:38:43.093Z</atom:updated>
            <content:encoded><![CDATA[<p>Hello everyone!</p><blockquote>This post is dedicated to all those people who are struggling to make a switch at software developer role.</blockquote><h4>Quick Intro</h4><p>I am Mohita Parashar, and I have been working in SAP Labs for approximately 3 years now. I would be joining Microsoft in the coming week.</p><h4>Background</h4><p>Started my career with a development support role.<br>When I realised, I spend more time wondering when the fruit basket in front of my desk will be restocked, than doing my work, I knew it was time for a change. 😂<br>I grabbed a 6 month fellowship opportunity to work on product management and development. I absolutely loved what I did! <br>So, I changed my team and became a full stack developer. For the past 1 year, I have been a full stack developer @ SAP AI Foundation team.</p><h3>How I Prepared for my interviews</h3><h4>Mission #1 : The one where I cracked the crooked Coding round</h4><h4>Level beginner</h4><p>I started giving interviews when I had gone through a fairly good number of coding questions tagged with companies Amazon and Microsoft on <a href="https://practice.geeksforgeeks.org/explore/?company%5B%5D=Amazon&amp;company%5B%5D=Microsoft&amp;page=1&amp;company%5B%5D=Amazon&amp;company%5B%5D=Microsoft">Geeksforgeeks</a> .</p><p>This list of questions is what I call “Standard Coding Questions”. This list almost covers 90% of the questions that were asked to me in my interviews at multiple companies including Amazon,Walmart, Cisco etc.</p><p>How to work this through?<br>1. Open this <a href="https://practice.geeksforgeeks.org/explore/?company%5B%5D=Amazon&amp;company%5B%5D=Microsoft&amp;page=1&amp;company%5B%5D=Amazon&amp;company%5B%5D=Microsoft">link</a>.<br>2. On the left, you should be able to see Filter and Sorting options.<br>3. One by one, go through all the questions in the following topic tags. Use the sorting option, to go through the easy questions first, followed by medium and lastly the hard ones.</p><pre><em>Binary Trees<br>Binary Search Trees<br>Linked List<br>Stack <br>Queue and Deque<br>Graph<br>Heap<br>Backtracking<br>Dynamic Programming.</em></pre><p>4. You should be able to finish this List in 30 - 45 days and Voila ! You’ll be ready to “START” giving the Coding rounds in most companies.</p><h4>Level Pro</h4><p>The above list will help you clear the coding rounds of most companies.<br>But there are some companies, in which you’ll be asked these standard questions, <em>with a twist</em>. <br>And just going through Geeksforgeeks is not sufficient to solve them in 30 – 40 mins.</p><p>So how to work this through ?</p><p>Code yourself!</p><p>Open up Leetcode . Solve its <a href="https://leetcode.com/problemset/algorithms/">Algorithms</a> section, topic wise, going easy to hard.</p><p>Solve the monthly Leetcode challenge everyday.<br>Keep a track of all the tricks and algorithms you used and REVISE!</p><p>You’ll get better each passing day.</p><h3>Mission #2 : The one where I demystified the Design round</h3><h4>Level Beginner</h4><p>Go through the HLDs and LLDs @ <a href="https://www.educative.io/courses/grokking-the-system-design-interview/B8nMkqBWONo">Educative.io</a> .</p><h4>Level Pro</h4><p>There are amazing videos on youtube. Watch these !</p><pre><a href="https://www.youtube.com/watch?v=xpDnVSmNFX0&amp;list=PLMCXHnjXnTnvo6alSjVkgxV-VH6EPyvoX">Gaurav Sen</a><br><a href="https://www.youtube.com/channel/UCZEfiXy7PmtVTezYUvc4zZw">CodeKarle</a><br><a href="https://www.youtube.com/channel/UCn1XnDWhsLS5URXTi5wtFTA">Tech Dummies</a><br><a href="https://www.youtube.com/channel/UC9vLsnF6QPYuH51njmIooCQ">System Design Interviews</a></pre><h3>Mission #3 : The one where they ask the basics</h3><p>While most companies will include one/more of these topics in technical rounds as well, some companies have dedicated rounds to test you on —</p><pre>OOPs <br>Multithreading and Concurrency<br>Computer Science Fundamentals (OS, CN, DBMS)<br>Language specific Knowledge<br>Leadership Principles (Amazon)<br>Architecture of your current project</pre><p>Geeksforgeeks is a nice resource to prepare most of these.</p><h3>Mission #4 : The one where I got rejected</h3><p>Getting rejected is an essential part of the job switching process. I have been rejected by 6 companies straight.</p><p>I was rejected when my interviews went bad, I was rejected even when I answered all questions, and I was rejected even when the interviewer said, “Hey this went good!”.</p><p>The trick is to get up every time. Learn from your mistakes. Take feedbacks from the interviewers.</p><p>Someday everything will make sense! So for now, laugh at the confusions, smile through the tears and keep reminding yourself, everything happens for a reason. 😌</p><h3>Mission #5: The one where you THINK you know it all now</h3><p>If you’ve been giving interviews for a while now, you’ll be well versed with all the topics I mentioned above, and you feel you have nothing more to study.</p><p>Give mock interviews. There are multiple websites which have this feature, like InterviewBit and Leetcode. Find people who are preparing for the same position as you are, and schedule mock interviews.</p><p>All the best!</p><blockquote>May the Source be with you ! <br> - My best friend</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4f9c2804b5f7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Metaflow]]></title>
            <link>https://mohitaparashar.medium.com/metaflow-97639d0bbddb?source=rss-7b1d59d20040------2</link>
            <guid isPermaLink="false">https://medium.com/p/97639d0bbddb</guid>
            <category><![CDATA[netflix]]></category>
            <category><![CDATA[metaflow]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[mlops]]></category>
            <dc:creator><![CDATA[Mohita Parashar]]></dc:creator>
            <pubDate>Mon, 06 Jul 2020 19:44:15 GMT</pubDate>
            <atom:updated>2020-07-07T03:53:20.851Z</atom:updated>
            <content:encoded><![CDATA[<h3>Metaflow : A boon for data scientists !</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*6L8uLoxmcwl4DhLV" /><figcaption>A data scientist’s condition</figcaption></figure><p>Soumya is a data scientist. She has to solve important business problems. To start, somebody will provide her with the business metrics ,that needs to be optimized. So, Soumya will -</p><p>1. Spin a notebook, gather data, slice and dice the data, get to number crunching , and figure out various stuffs and ways about how to optimize the business metric.</p><p>2. Use packages like TensorFlow, and write her own training code.</p><p>3. All of these codes, need to run on some data set. So she will use some ETL , so that she can feed the data to her model.</p><p>4. Production workflows should run on servers, not on a laptop. So she will schedule these model somewhere, like cloud.</p><p>5. These days, servers execute containers. Workflows need to be containerized for execution.</p><p>6. Think where should the results go? Sometimes they are deployed as containers too.</p><p>7. Finally containerized models can be consumed by business applications.</p><p>The stakeholders evaluate the results, and they want more models , better features, more demands and so on, thus requiring Soumya to repeat the entire process again.</p><p>So how much time did Soumya actually spend on Data Science? Pretty minimal! Most of her time is spent on writing ETL, figuring out where to get the data from, scheduling the compute, and if anything fails, it’s very difficult to triage what happened!</p><p>Data science is a very iterative process, and generally it is guaranteed that you’d not succeed in your very first attempt. There’ll be a plethora of models Soumya would be working with. So she needs to keep track of the different experiments that she had launched, and monitor which ones are working, which ones are not! So her another job is experiment management and tracking.</p><p>That’s a lot of work for a single data scientist!!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*fe83oZQVip6ncWlX.png" /><figcaption>A general ML stack</figcaption></figure><p>That’s when Metaflow comes in. What metaflow does is, provide an unified abstraction to this ML stack, so that Soumya can be more of data scientist and less of a system engineer.</p><blockquote>Metaflow provides a unified API to the infrastructure stack that is required to execute data science projects, from prototype to production.</blockquote><p>Let’s see how that happens.</p><p>So this what any ML code at the end of the day looks like.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uwSFQDIpnp4OQ9Z5IS6vDQ.png" /></figure><p>What metaflow does is, it allows you to declare your computation as a Directed Acyclic Graph(DAG).We call the graph of operations <strong>a flow</strong>. You define the operations, called <strong>steps. </strong>Steps become nodes of the graph and contain transitions to the next steps, which serve as edges. You can construct the graph using an arbitrary combination of the following three types of transitions supported by Metaflow:</p><p><strong>1.Linear</strong></p><p>The most basic type of transition is <strong>a linear</strong> transition. It moves from one step to another one.</p><p>Here is a graph with two linear transitions:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/389/0*CyTGpfO_hn1juHrI" /></figure><p>The corresponding Metaflow script looks like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/661/1*VuDH8h3_-TSJ3rR6MgyvBw.png" /></figure><p>This flow creates <strong>a data artifact</strong> called my_var. In Metaflow, data artifacts are created simply by assigning values to instance variables like my_var.</p><p><strong>2.Branch</strong></p><p>You can express parallel steps with <strong>a branch</strong>. In the figure below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/529/0*BGRiSACRN8a12U5k" /></figure><p>Any number of parallel steps are allowed. A benefit of a branch like this is performance: Metaflow can execute a and b over multiple CPU cores or over multiple instances in the cloud.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/611/1*N-i8UvtCR9DU_MCCpqglWw.png" /></figure><p><strong>3. Foreach</strong></p><p>The above transition is useful only if you know the branches already at development time. But what if you want to branch dynamically based on the data. This is the usecase for Foreach branch.</p><p>Foreach works similarly like the branch above but instead of creating named step methods, many parallel copies of steps inside a foreach loop are executed.</p><p>A foreach loop can iterate over any list like titles below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/554/0*bzOoyvdyAfo77AAz" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/565/1*3Vl6Ax0PizSKJgBxGiI9Yg.png" /></figure><p>Steps inside a foreach loop create separate <strong>tasks</strong> to process each item of the list. Here, Metaflow creates three parallel tasks for the step a to process the three items of the titles list in parallel.</p><h3><strong>Features:</strong></h3><ol><li>Automatic State transfer:</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/326/1*367-zdo98KUkRRsZgoSWvA.png" /></figure><p>In tools like AirFlow, the nodes are isolated units of computation. The state transfer is left upto the users, leaving them confused with what to pass to the other containers and what to store and what not to.</p><p>And that’s what metaflow does.<br>Every single time, your step finishes, it snapshots the entire state and it is magically transferred to all the subsequent steps. So at any point of time, if something fails, we can come back and check what the state of our workflow was in a particular stage.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/537/1*TcmlZ4qG_1NUJONNTfsMHw.png" /></figure><p>2. Resuming a run :</p><p>The fact is, storage is cheap, so when things fail, you need to make sure you have access to the tracking history and you feel at that time, it would have been a good idea if you could have recorded the entire world at that time!</p><p>As a result of state transfer, if a run failed (or was stopped intentionally)and you fixed the error in your code in a particular stage, you could restart the workflow from the stage where it failed/stopped.</p><p>3. Versioning:</p><p>Every time you execute a workflow, a dedicated namespace is created for that particular user and are allocated run ids, and then bundled in a client API.</p><p>Metaflow persists all runs and all the data artifacts they produce. Every run gets a unique run ID, e.g. HelloFlow/546, which can be used to refer to a specific set of results. You can access these results with the Client API.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/686/0*ynByi_N4XEyLE4hU" /></figure><p>4. Real time Monitoring</p><p>You can monitor status of your workflow in a Jupyter notebook abstraction.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/549/1*htf4nn-zAsFPFVOp3vVTaw.png" /></figure><p>5. Vertical Scalability</p><p>In this world of cloud, we have access to incredible resources on our finger tips. But our laptops have limited capacity and more often , the data sets are in petabytes, needing 8 GPUs. What do you do then?</p><p>You wish you package that piece of code and deploy and run it on specific resources. But this is difficult even for an engineer, imagine about data scientists!</p><p>Metaflow relies on decorators and we let us specify the resources we need .<br>Your function then gets packaged and is run on top of a docker container on an instance that has these resources.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/575/1*OJaIzY6H0cST8No5dOpZFQ.png" /></figure><p>6. Horizontal Scalability:</p><p>Users can parallelize the tasks using Foreach.</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/0*aAr4M2rafLjwefU_" /></figure><p>7. Dependency Management/ Reproducibility:</p><p>Reproducibility is a core value of Machine Learning Infrastructure. It is hard to collaborate on data science projects effectively without being able to reproduce past results reliably. Metaflow tries to solve several questions related to reproducible research, principle amongst them, dependency management: how can you, the data scientist, specify libraries that your code needs so that the results are reproducible?</p><p>Metaflow aims at solving both the questions at once: how can we handle dependencies so that the results are reproducible? Specifically, it addresses the following three issues:</p><ol><li>How to make external dependencies available locally during development?</li><li>How to execute code remotely on AWS Batch with external dependencies?</li><li>How to ensure that anyone can reproduce past results even months later?</li></ol><p>Metaflow provides an execution environment context, --environment=conda, which runs every step in a separate environment that only contains dependencies that are explicitly listed as requirements for that step. The solution relies on <a href="https://conda.io/docs/">Conda</a>, a language-agnostic open-source package manager by the authors of Numpy.</p><p>$ python LinearFlow.py — environment=conda run</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/638/1*kDVg2P-owOZLBcNTB4O3GA.png" /></figure><p>8. Sandbox-</p><p>Metaflow comes with built-in integrations to various services on AWS. The seamless integration to the cloud is a key benefit of Metaflow.</p><h3>Downsides:</h3><ol><li>Tightly coupled with AWS.</li><li>Not supported on Windows.</li></ol><p>And that’s all! While Metaflow seems to be a step in the right direction, there’s still a long way to go until it is ready for universal use. Thank you for reading!</p><p>References:<br>1.<a href="https://docs.metaflow.org/">https://docs.metaflow.org/</a><br>2. <a href="https://www.youtube.com/watch?v=-oMZAS9qfrE">More Data Science, Less Engineering with Netflix’s Metaflow By Savin Goyal</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=97639d0bbddb" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>