<?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 Devang Mundhra on Medium]]></title>
        <description><![CDATA[Stories by Devang Mundhra on Medium]]></description>
        <link>https://medium.com/@devangmundhra?source=rss-44aef2fce762------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*V5LIcfcwfV0GWHOEbobdAg.jpeg</url>
            <title>Stories by Devang Mundhra on Medium</title>
            <link>https://medium.com/@devangmundhra?source=rss-44aef2fce762------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 22:11:33 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@devangmundhra/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[Banking : Winds of Change]]></title>
            <link>https://medium.com/@devangmundhra/banking-winds-of-change-f8534397d9da?source=rss-44aef2fce762------2</link>
            <guid isPermaLink="false">https://medium.com/p/f8534397d9da</guid>
            <category><![CDATA[banking]]></category>
            <category><![CDATA[fintech]]></category>
            <dc:creator><![CDATA[Devang Mundhra]]></dc:creator>
            <pubDate>Mon, 20 Jul 2020 06:51:53 GMT</pubDate>
            <atom:updated>2020-07-20T06:51:53.126Z</atom:updated>
            <content:encoded><![CDATA[<h3>Banking : Winds of Change</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/886/1*U1b8DC6JCcr6NqQ09IgF8g.jpeg" /></figure><p><em>Note: This article has been written using references mostly from India from where I am based, but similar trends are there across different geographies with some regions like Europe, UK and Australia well ahead in their implementation and other regions catching up soon</em></p><p>Banks are interesting institutions in that almost anything can impact them — changes in technology, economy, fiscal actions, regulations, data privacy, security markets, trade deals, climate change, cyber security, customer expectations — and such. And yet (or maybe because of that?), banks are some of the most enduring institutions in an economy.</p><p>All factors remaining constant, this article provides a couple of macro trends that banks might be seeing today given the changes in competition, regulation and customer expectations, and how they can adapt. While many of the other factors mentioned above might themselves need a much longer discussion and speculation, I have kept the scope of this article very limited.</p><h3>Headwinds</h3><h4>Declining Distribution</h4><p>A long-running trend across industries, geographies and over the times have been people and organizations increasing their specialization and becoming more narrow in their focus. One big reason for this trend has been consumer needs becoming more diverse and consumers themselves becoming more demanding from their service providers giving rise to <em>nichification </em>of products and experiences.</p><p>Banks however have bucked this trend and remained similar in their form and function over time. While the origin story of different banks could be stemming from a once-held niche, they all converge to similar end states. Some reasons could be that regulators prefer standardization, or that banks grow by hiring people from other banks so end up having similar views and modes of operations.</p><p>Banks are also keenly aware of this, and are trying a variety of things to improve customer experience — from converting <a href="https://www.accenture.com/_acnmedia/PDF-91/Accenture-Bank-Branch-Digital-Disruption-from-Transactions-to-Conversations.pdf">branches to lounges</a>, to coming up with multiple versions of the same product to cater to different customer segment. However, these efforts will be running into headwinds of <em>banks becoming less of a destination and more transactional</em>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/348/1*kmeAga3zA0BC4ACC1ea6sA.gif" /><figcaption>Different(14) versions of current accounts by India’s largest Private Bank. Consumers are looking for more curated experiences per their needs. Source: hdfcbank.com</figcaption></figure><p>The new destinations are the companies and apps that are focusing on single use case (niche) problems. While banks will keep trying to address changing customer expectation, on the distribution front they will start competing with products who are solving very specific problems, like crop price discovery for a farmer, new demand generation for a fleet owner, or CRM for a solopreneur. Many of these use cases have a financial transaction as part of their workflow and this is where these apps can better distribute the right product at the relevant time in a seamless manner compared to the banks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/635/1*rpjuLWXYC_8fi7AiYebDfQ.jpeg" /><figcaption>2009 Apple Ad — There’s an App For That. Source: <a href="https://www.cultofmac.com/62892/apple-gets-a-trademark-theres-an-app-for-that/">CultOfMac</a></figcaption></figure><p>Not only that, banks are going to be at an inherent disadvantage when looking to distribute products on their own since these distribution app companies have the ability to spend much more for customer acquisition. They can not only employ more techniques for acquisition (risk capital infused inorganic growth, network effects from product usage, referral incentives, novelty factor), they can also spend more on CAC since their revenues will depend on other sources of income as well like charging for convenience, selling data to advertisers, and such. Banks typically don’t (and in most jurisdictions can’t) do that.</p><h4>Brand Irrelevance</h4><p>Since banks are not owning the distribution, the part of the transaction they are involved in is being increasingly getting commoditized. Technology companies <a href="http://www3.weforum.org/docs/Beyond_Fintech_-_A_Pragmatic_Assessment_of_Disruptive_Potential_in_Financial_Services.pdf">keep pushing on this front </a>— payment gateways in payments, o<a href="https://en.wikipedia.org/wiki/Open_banking">pen banking</a> platforms commoditizing integrations and Data-As-A-Service companies providing transactional data. At the same time regulatory trends like <a href="https://www.npci.org.in/product-overview/upi-product-overview">UPI</a>, <a href="https://en.wikipedia.org/wiki/Payment_Services_Directive">PSD2</a>, Account Aggregators, and other such initiatives are further forcing the banks to be interoperable. This time is almost reminiscent of the telecom industry when number portability reduced the friction to switch from one service provider to another, and a lot of legacy companies with no differentiators started to see declines.</p><p>In addition, with number of banks increasing (<a href="https://m.rbi.org.in/commonman/english/scripts/banksinindia.aspx">more than 100 banks i</a>n India) and each bank coming up with many versions of the same product, the discovery of these products are increasingly being done on aggregator sites who provide better range, user experience, discoverability and promise of finding the best product that suits the requirements. Just like with rise of more news sites on the internet gave prominence to search engines and news aggregator and publications started to lose prominence as stand alone brands, similarly product manufacturers might start to lose prominence as the product becomes commonplace.</p><h3>Tailwinds</h3><p>In these changing landscapes, banks then need to double down on the strengths that they have — strengthening relationships and providing unique and custom solutions for large customers, and becoming an infrastructure provider to the companies who are doing a better job at distributing the financial products.</p><h4>Relationships</h4><p>One segment where the app companies have not found a differentiating distribution model are mid and large corporates. Large enterprises often have very different and unique requirements for a generic product to be built and distributed at scale, and at the same time the distribution process is very similar — through enterprise sales, problem discovery and relationship building. So unlike in other segments, banks don’t have any inherent disadvantage to cater to large companies directly and will in fact have advantage in terms of large companies being more comfortable with banks directly. Similarly, banks should partner with the app companies to help them provide a better and more efficient way to distribute their products.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*u5nD07ge7_1mP2T9xZJDtw.png" /><figcaption>Large Industry Credit is already in the top-2 sectors the banks operate in. Source: <a href="https://tijorifinance.com/sector/banking#opmetrics-sector">TijoriFinance</a></figcaption></figure><h4>Infrastructure</h4><p>For the other customer segments — small businesses, retail consumers and such — banks can become infrastructure providers to the apps who are serving to these customers rather than competing with the apps for distribution. The apps would want to focus on their use cases and not worry about the regulatory, compliance and all the components of the financial network.</p><p>So far, banks to a large extent, have seen most partners as lead-gen opportunities trying to control customer experience once a lead is acquired. This will flip with the distributors partnering with banks to use their infrastructure. Here, the banks who have the best scale, integration possibilities and are developer focus will win. Just like AWS got a head start with providing cloud based infrastructure with other companies playing catch-up, those banks who can adapt early will benefit.</p><p>So bottom line, app companies have a better ability and visibility to solve specific problems of their target customers and are comfortable building and distributing niche solutions. Some of these solutions have financial transactions as part of their workflow. Financial service providers should start to look at themselves as finance infrastructure companies and invest heavily today to get a pole position. 🏁</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f8534397d9da" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[5 Great Tools That Didn’t Work Out — Software Project Management]]></title>
            <link>https://medium.com/kredx-engineering/5-great-tools-that-didnt-work-out-software-project-management-667563c513b?source=rss-44aef2fce762------2</link>
            <guid isPermaLink="false">https://medium.com/p/667563c513b</guid>
            <category><![CDATA[project-management]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[engineering-management]]></category>
            <dc:creator><![CDATA[Devang Mundhra]]></dc:creator>
            <pubDate>Tue, 23 Jul 2019 09:56:01 GMT</pubDate>
            <atom:updated>2019-07-23T09:56:01.213Z</atom:updated>
            <cc:license>http://creativecommons.org/publicdomain/zero/1.0/</cc:license>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qmC1dhOEMw_4awMaT-Uf8w.jpeg" /><figcaption>What always starts on a whiteboard…</figcaption></figure><h3>5 Great Tools That Didn’t Work Out — Software Project Management</h3><p>Over the years we have tried many things at <a href="http://www.kredx.com/">KredX</a> to manage estimates and the delivery of tech and product projects. We tried some common tools and some not-so-common-ones, and still, our search continues.</p><h3>Requirements</h3><p>We are a small team of engineers and PMs — about 20–25 people in all. And we just have a short ask from the project management tool:</p><ol><li><strong>Super-easy</strong> to enter tasks (including future, ad hoc and current tasks) — It should take less than 5 seconds to enter a task, or better still, to somehow get that input automagically.</li><li><strong>Super-fast </strong>to add and update attributes — owner, projected delivery date, which project is it a part of, status, time estimate and actual completion date should be a non-thought to do.</li><li><strong>Cost </strong>under $7 per user per month (more useful than Slack? — prove it).</li><li><strong>**Reports**</strong> available for all users to show what was estimated, what was delivered, the reason for the difference, and improvements over time.</li><li><strong>Design</strong> to want to use it.</li></ol><p>There are a few other good-to-have things like ticket tracker for external teams, dependency management, etc. but that can be skipped for now. So that’s it, nothing fancy.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fgiphy.com%2Fembed%2FSTLj8QUhg25pe%2Ftwitter%2Fiframe&amp;display_name=Giphy&amp;url=https%3A%2F%2Fgiphy.com%2Fgifs%2Fmrw-describe-asked-STLj8QUhg25pe&amp;image=https%3A%2F%2Fmedia.giphy.com%2Fmedia%2FSTLj8QUhg25pe%2Fgiphy.gif&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=giphy" width="435" height="195" frameborder="0" scrolling="no"><a href="https://medium.com/media/c1c30263ca53b176aa1d41ac6f487f19/href">https://medium.com/media/c1c30263ca53b176aa1d41ac6f487f19/href</a></iframe><h3>Tools and trials</h3><p>We have already tried many tools for this purpose, using each for at least 4 months before calling it quits on the tool. Here is our experience so far:</p><h4>1. Asana [Nov 2016 — March 2017]</h4><p>Asana was the hot tool a couple of years ago for project management and we dove right in. A lot might have changed by now, I assume, but at that time it seemed overly complex (at least 4–5 clicks and form entries) to add tasks, figure out statuses of a project quickly, and the lack of reports for feedback and improvement was a downer.</p><h4>2. Google Sheets + Github Issues + Trello + YouTrack [March 2017 — September 2017]</h4><p>Post Asana, we went into a discovery phase with each engineer/PM championing their preferred tool. We had long debates about which tool is better, building different workflows for the different tools, but never felt like we had found “The One”.</p><h4>3. Jira [September 2017 — August 2018]</h4><p>Then we found an ambassador for Jira who was willing to take up the ownership of getting everyone to use it and make sure that all the entries were properly tagged, and new users were set up properly. It worked to a certain extent, but then we finally gave up because</p><ol><li>Even after a few months, the team depended on the ambassador to keep the tool working fine</li><li>It had become a bit expensive</li><li>It was still difficult to input data to create tasks</li><li>A dearth of reporting</li></ol><h4>4. Monday.com [August 2018 — March 2019]</h4><p>Then someone in the team suggested this new tool and we thought, sure why not! It was very responsive, and allowed us to customize it exactly the way we wanted…well…to a certain extent. We started to see limitations with all the related data being required to be at the same place which caused the pages to slow down, and reporting not being available, as mentioned in Requirement #4.</p><h4>5. Airtable [March 2019 — Present]</h4><p>Which brings us to the current solution - a spreadsheet with foreign keys. Similar to Monday.com — we customized it substantially to work for our team, added a bunch of formulae, pointers and color-coding to make it easier to digest. It is a hacky approach but more or less works for us. For reports, we just download the CSV and use Excel for the charting. Default options would save a lot of data entry.</p><p>I feel very confident that we didn’t use these tools properly as they were supposed to be used. After all, everyone else seems to be going about project management just fine. That is what I thought until I spoke to a few tech and product leaders and realized many have a similar story like ours to share. Many teams have figured it out, while others still continue to search.</p><p><strong>Got thoughts on this subject? Please do share your journey (and destination) in comments, or on your blog and post a link here.</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=667563c513b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/kredx-engineering/5-great-tools-that-didnt-work-out-software-project-management-667563c513b">5 Great Tools That Didn’t Work Out — Software Project Management</a> was originally published in <a href="https://medium.com/kredx-engineering">KredX Engineering</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[BLR — BOM — BLR [20–21 Aug ‘18]]]></title>
            <link>https://medium.com/@devangmundhra/blr-bom-blr-20-21-aug-18-5ac1edbd3e9d?source=rss-44aef2fce762------2</link>
            <guid isPermaLink="false">https://medium.com/p/5ac1edbd3e9d</guid>
            <category><![CDATA[sister-love]]></category>
            <category><![CDATA[travel]]></category>
            <category><![CDATA[personal]]></category>
            <category><![CDATA[late]]></category>
            <dc:creator><![CDATA[Devang Mundhra]]></dc:creator>
            <pubDate>Sat, 25 Aug 2018 08:50:27 GMT</pubDate>
            <atom:updated>2018-08-25T09:02:19.836Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DnQsEwqQUGIQ3pqR4lsMfg.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/photos/BjcGdM-mjL0?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Andy Beales</a> on <a href="https://unsplash.com/search/photos/luggage-man-run-airport?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>There are days when things go well. And then there are other days.</p><h4>Onward</h4><p>I am a person who usually likes to time things well - to save a few minutes here and a few there. Whether it is using all four stoves to make food in parallel, or listening to podcast at 1.5x. One of my prized possessions is an automatic watch that I had got years ago, a constant reminder to me on how physics and ingenuity can be creatively combines to produce pieces of arts. Though the watch has started to fall behind a few minutes every couple of days, I still use it and keep winding it (watches tend to accumulate sentimental values over time).</p><p>So as I was saying, I am a person who likes to time things well — catch up a few minutes more with friends and then be rushing to catch a flight or a train. Yup, the person you may have seen at the airport often sprinting towards security and requesting you — strike that — begging you to let him go in the security line ahead of you: thats me. I don’t like doing that, my wife hates me for doing this, but like a promise of never again after a hangover, it happens again.</p><p>But not on 20th August 2018. There was a morning meeting in Mumbai on 21st, so I left from office for Bengaluru Airport well in time, and reached at 3pm for a 4:35pm flight. Comfortably checked in and security check done, I still had about 1hr to kill. So after catching up on some work, I went over to the Relay bookshop to browse through books and get a book for the flight.</p><p>Choosing a book is tough. There are so many options — and then there are all these other factors to weigh in: your interests, judgements of others (kindle often helps here), size of the book (yup, I have book commitment issues). So went through the Top 10, books on sale, fiction, non-fiction, etc. Flipped through a couple of books, and finally decided on APJ Abdul Kalam’s <em>My India — Ideas for the future </em>for being a light read. I wanted to pick up a thriller book as well for the journey but I saw the time on my watch that it was already 4pm so got the billing done and rushed to the gate. Found that the gate was on a different level, so hopped through the stairs to see the gate closing. Reporting at gate for GoAir flight — 4:14pm. Time on my watch — 4:07pm. Oops.</p><p>There were already some passengers arguing with the onground GoAir staff, requesting them to let them go but nope, GoAir was extremely strict for being late even by a minute. Long story short, they got me out of the airport again and said — “Sir, we can get you on tomorrow’s 4:35pm flight, or you need to figure out your own way to Mumbai”.</p><p>So tomorrow 4:35pm was not an option. I checked with other airlines if there were tickets available to Mumbai for the evening. Nope — all sold out except an evening ticket for ~Rs. 65K and a late night ticket for ~Rs. 35K 😳. Smiling, complimenting and requesting the GoAir duty manager didn’t help. (Though kudos to the duty manager, to face constant ire of customers day in and day out shouting in your face. Hard job.). So after some thinking and checking with office to make sure tomorrow’s meeting wasn’t cancelled, decided that the reasonable option would be to go to Pune and take a cab to Mumbai after landing.</p><p>Got a Jet flight to Pune for 21:45. Cool — got some time to spend. Went to the lounge, had some food and finished some calls. Checking time constantly on the internet synchronised phone time. No chances this time.</p><p>Flight delayed 🤬. Now landing at 00:30. Started looking for a hotel near the airport in Pune, and simultaneously arranging for a cab for 5am from Pune to Mumbai. Booked.</p><h4>Stay</h4><p>Woke up at 4:30am, got ready and got in the cab to Mumbai. It had been raining all night and the road was quite foggy. Between trying to catch up on sleep and talking to the driver, the 3hr journey was a blur. Had breakfast, ready for meeting. Meeting went as meetings go.</p><p>Borrowed a car to drive to a brother’s place. Was driving with Google Maps on, and being a borrowed car, without bluetooth connectivity. Wife called, so (out of love/fear ?) picked it up and started chatting. And voila, two policemen pass me and ask me to stop the car at the side, and asked for my driving licence.</p><blockquote>Policeman (P) — Sir, you were talking on the phone?</blockquote><p>Me: Yes sir.</p><blockquote>P: (surprised) You’re breaking the rule.</blockquote><p>Me: Really? I am sorry, didn’t know Mumbai had this rule. I have just come from Bangalore today.</p><blockquote>P: (confused) — Sir, this is a rule all over India.</blockquote><p>Me: Really, I had no idea! No one in Bangalore has ever told me that, and given the traffic conditions there, I am often on the phone.</p><p><em>[In all honesty, I had an inkling that this must be a rule, and it’s safer anyways for distraction free driving, but I haven’t seen TV for a long time and there are no posts talking about this rule on the road, so it slipped my mind].</em></p><blockquote>P: The fine is Rs. 2500. Should I give you a paper challan or e-challan.</blockquote><p><em>Ok — many things to process here. No trying to negotiate an off-the-book transaction. Joy swelling. Rs. 2500! Grief building. Options for challan? Intriguing.</em></p><p>Me: Whats the difference between the challan?</p><blockquote>P: Paper challan — it’s Rs. 2500 but you’ll get your licence back. e-Challan — you’ll have to pay Rs. 200 but you’ll get your licence later.</blockquote><p>Me: (My time to get confused now) So if I pay e-challan, when will I get the driving licence back? Note that the address on the licence is of Vadodara, I live in Bangalore and we’re in Mumbai, but I’d want the licence delivered in Bangalore.</p><blockquote>P: So you want to pay by echallan?</blockquote><p>Me: If I pay by e-challan, I just need to pay 200? Nothing after that?</p><blockquote>P: Echallan you can only pay by credit card at the next signal.</blockquote><p>Me: So by cash its Rs. 2500 and e-challan is Rs. 200? So much difference!? How come?</p><blockquote>P: Yes, but with cash you’ll get the licence back immediately and you’ll get a receipt. Court will refund half of the money back.</blockquote><p>Me: Ok — so in e-challan when and where will I get my licence?</p><blockquote>P: (A bit irritated now) I have to talk to this other person I have stopped, why don’t you think about this and let me know what you want to do about this.</blockquote><p>Me: Ok, but….</p><p><strong><em>(2 minutes later)</em></strong></p><blockquote>P: So, what is it going to be?</blockquote><p>Me: Sir, why not give me a warning and let me go this time. I really didn’t know about the enforcement of the rule.</p><blockquote>P: No no, you paying the money is going to be a warning. Otherwise, next time, you’ll say the something to someone else that please give me a warning.</blockquote><p>Me: …</p><blockquote>P: …</blockquote><blockquote>P: So what is it going to be? Challan or echallan?</blockquote><p>Me: I am still not clear about the difference, why would anyone pay Rs. 2500 if they can pay 200?</p><blockquote>P: (Giving up) Ok — I’ll give you a warning this time. But don’t speak on the phone next time.</blockquote><p>Me: Yes sir, definitely. Thank you!</p><p><em>Mumbai Police — Respect!</em>🙏🏻</p><p>Finally reach brother’s place.</p><h4>Return</h4><p>So with dinner done, sister takes me to the airport. Got one laptop bag and one small carry-on baggage. This is Western Expressway on a rainy day in Mumbai, and I didn’t want to take any chances. Enough of mis-adventures on this short trip already. For a 9:30pm flight, leave at 7:50pm with Google Maps showing 40mins. 30 mins of buffer.</p><p><strong>7:50pm</strong>: Sister didn’t allow me to take a cab, and insisted she drive me to the airport. No chances this time.</p><p><strong>8:10pm</strong>: 20 mins on the road. Google Maps showing 36 mins remaining. Sis getting agitated.</p><p><strong>8:30pm</strong>: Another 20 mins on the road, Google Maps showing 26 mins remaining. Sister hyperventilating. Estimated time to arrival — 20mins before departure.</p><p>Flight is a sure miss if this trend continues. I left the carry-on bag in the car, got off in the middle of the road, stopped a motor cyclist and offered 200 bucks to drop to the airport. He was a Domino’s delivery boy, and I sat on the delivery scaffolding on the bike. Snaking through the traffic, overtaking cars slowly, approaching airport steadily.</p><p><strong>9:05pm: </strong>Reached airport. Rushed through the security — begging people to let me go through the security line ahead of them. On the other side of the security Indigo ground crew jumping up and down, asking me to come fast. Finally went through the security, things falling from my laptop bag and the Indigo crew helping me pick up the mess I was leaving behind. And finally, just manage to get into the bus to the plane. <em>Phew</em>.😅</p><p><strong>9:25pm: </strong>All boarded and plane locked up, pilot announces we will have to wait in the runway for 40mins because of air-traffic congestion.</p><p>F.M.L. 😑</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5ac1edbd3e9d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Deep Learning: Highlights of 2017]]></title>
            <link>https://medium.com/@devangmundhra/deep-learning-highlights-of-2017-adea2ed6319f?source=rss-44aef2fce762------2</link>
            <guid isPermaLink="false">https://medium.com/p/adea2ed6319f</guid>
            <category><![CDATA[neural-networks]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Devang Mundhra]]></dc:creator>
            <pubDate>Sun, 07 Jan 2018 13:30:15 GMT</pubDate>
            <atom:updated>2018-01-09T06:52:14.319Z</atom:updated>
            <cc:license>http://creativecommons.org/publicdomain/zero/1.0/</cc:license>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*y2XLwBgjGqfRGG6AH6iS7A.png" /><figcaption>A classification problem (Source: Wooden Shape Sorter, Bannor Toys)</figcaption></figure><p>Deep Learning continues to have an amazing streak. And just like GoT’s delayed finale, winter is not coming here anytime soon 😀. Probably the only other technology last year that has surpassed deep learning in terms of hype and excitement has been blockchains — specifically because of its use in cryptocurrencies. However, while blockchains came into the pop culture because of speculation on cryptos, deep learning was interesting because of new ingenious models being built, and the expansion of capabilities on what these learning networks are able to do.</p><p>Some highlights of deep learning in 2017.</p><ul><li><em>Neural Networks got better in their skills</em> — whether it was better <a href="https://www.microsoft.com/en-us/research/blog/microsoft-researchers-achieve-new-conversational-speech-recognition-milestone/">speech-to-text</a>, playing <a href="https://deepmind.com/blog/alphago-zero-learning-scratch/">Go</a> or <a href="https://blog.openai.com/dota-2/">DOTA</a>, <a href="https://www.theverge.com/2017/7/12/15957844/ai-fake-video-audio-speech-obama">voice modulation</a> and tons of other things.</li><li><em>Computers exploring their creative sides</em> — Last year was a bit of a renaissance for deep neural networks in which they continued to find their own creative edges with applications like <a href="https://deepmind.com/blog/wavenet-generative-model-raw-audio/">writing music pieces by themselves</a> and <a href="https://blog.openai.com/robots-that-learn/">imitating others</a>. For a bit it seemed they even <a href="https://www.wired.com/story/facebooks-chatbots-will-not-take-over-the-world/">found their own language</a> to communicate!</li><li><em>Neural Networks’ Adversaries </em>— A bit more insights in what causes <a href="https://blog.openai.com/adversarial-example-research/">neural networks to get fooled</a> with small, very specific perturbations in training data. Also, <a href="https://medium.com/microsoft-design/how-to-recognize-exclusion-in-ai-ec2d6d89f850">bias in deep learning</a> became a more important topic to study given so many decisions are getting made by data models.</li><li><em>Peering into deep learning black box</em> — Some <a href="https://www.tensorflow.org/get_started/graph_viz">tools</a> were made available to help understand what happens when networks are trained over time, and how networks <a href="https://distill.pub/2017/feature-visualization/">see</a> the world. Also some attempts to understand the <a href="https://blog.acolyer.org/2017/11/24/on-the-information-bottleneck-theory-of-deep-learning/">information</a> <a href="https://blog.acolyer.org/2017/11/15/opening-the-black-box-of-deep-neural-networks-via-information-part-i/">theory</a> of deep learning, to try make designing networks bit more of a science than art.</li><li><em>Just the beginning</em> — Geoffrey Hinton and his team presented <a href="https://medium.com/ai%C2%B3-theory-practice-business/understanding-hintons-capsule-networks-part-i-intuition-b4b559d1159b">a new model for deep learning</a> that would help in processing images with much more accuracy and lesser data for training. He also suggested that we need to scrap how DL is done currently and try to <a href="https://www.axios.com/artificial-intelligence-pioneer-says-we-need-to-start-over-1513305524-f619efbd-9db0-4947-a9b2-7a4c310a28fe.html">start over again</a>!</li></ul><p>Even with all the advances and increasing optimism about the potential of AI, there are still some deep rooted limitations about what AI can do.</p><p>There is a <a href="https://arxiv.org/pdf/1801.00631.pdf">recent paper</a> from Gary Marcus which provides some insights into the current shortcomings of deep learning, where he critically appraises his own field. <em>TL;DR Deep learning is data hungry, and skills learned by one network, or any prior knowledge is difficult to transfer to others. It cannot extract deeper principles from the data (like hierarchies, causation vs. correlation) and are very much influenced by training sets.</em></p><p>One way to look at the current deep learning methods is that it is essentially a classification problem.</p><blockquote>Deep learning, as it is primarily used, is essentially a statistical technique for classifying patterns, based on sample data, using neural networks with multiple layers.</blockquote><p>The more diverse data one can provide to the problem, the better it can classify. Whether it is supervised or unsupervised, it identifies patterns in the data and tries to classify new data based on the patterns learnt. Which is possibly also the reason why deep learning algorithms work best when there is only a <em>finite</em> range of outcomes like classes in an image set, characters in a language or attributes in a music set— because classifying infinite possibilities is unviable!</p><p>2018 promises to be an exciting year for AI. All the big cloud companies are providing more and more ready to use services built on top of ML. Native AI capabilities are getting pushed to the edges including phones, which means a lot many applications are going to start using AI. Additionally, advances in hardware from tensor processing to qubits processing could help answer even more complex challenges found intractable even with today’s computing resources.</p><p>Thank you much <a href="https://blog.acolyer.org/">Morning Paper</a>, <a href="https://distill.pub/">Distill Pub</a> and <a href="https://blog.openai.com/">OpenAI Blog</a> for being a major source of information on what&#39;s happening in machine learning last year, and taking the effort to make it understandable for us mortals. And a deep gratitude for the countless enthusiasts and scientists who have shared their code and understandings for others to freely use and work on. You folks rock!🤘🏻</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=adea2ed6319f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A conversation with Engineering at EkStep]]></title>
            <link>https://medium.com/@devangmundhra/a-conversation-with-ekstep-engineering-2278e7b5f143?source=rss-44aef2fce762------2</link>
            <guid isPermaLink="false">https://medium.com/p/2278e7b5f143</guid>
            <category><![CDATA[remote-working]]></category>
            <category><![CDATA[engineering-mangement]]></category>
            <category><![CDATA[distributed-teams]]></category>
            <category><![CDATA[engineering]]></category>
            <dc:creator><![CDATA[Devang Mundhra]]></dc:creator>
            <pubDate>Sat, 25 Nov 2017 10:54:55 GMT</pubDate>
            <atom:updated>2017-11-26T02:00:30.850Z</atom:updated>
            <cc:license>http://creativecommons.org/licenses/by/4.0/</cc:license>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XHGmMAn4PmF78zzrxpAkQQ.jpeg" /><figcaption>Engineering through inclusion</figcaption></figure><p>Recently I had the opportunity to visit EkStep’s office to learn more about their mission, and how they are scaling engineering to keep pace with their ambitious mission. It was a fantastic experience to discuss engineering practices and tradeoffs of using different approaches with EkStep engineering team — <a href="https://www.linkedin.com/in/pramodkvarma/">Pramod</a> and <a href="https://www.linkedin.com/in/madhuchandrar/">Madhu</a> — and a<a href="http://primevp.in/#portfolio"> few other startup peers in Bangalore</a>. What was thought provoking for me was the paradigm shift on how EkStep approached building their engineering team.</p><p>For those not familiar with <a href="https://ekstep.org/">EkStep</a>, it is a not-for-profit organisation to provide learning opportunities to children who hitherto did not have the opportunity to learn. And this, they are doing by building an open, agile and scalable platform for anyone to plug-and-play solutions, content and ideas (think AWS for education)!</p><p>The team at EkStep understood that their strength is in building large, “societal”-scalable systems which can spawn and support an ecosystem of diverse stakeholders. Hence they are focusing on two things — building a platform for anyone to use, and fostering partnerships to build use-cases and adoption. The focus of our discussion was on the former — how to build a scalable, open and heterogeneous platform through collaboration of partnerships and not needing to hire people in-house. EkStep, under the technical leadership of Pramod and Madhu, have been able to foster partnerships with many organisations who allow their employees to work with EkStep and it’s other partners, and have been able to scale the approach to 100+ developers and architects working in a collaborative manner. It is almost like a semi open source model.</p><p>While the discussion covered a lot of aspects, following are eight points that came up in the discussion on how a development paradigm like this can be cultivated.</p><h4>Selection</h4><p>The first step is identifying the right partners, or rather the right individuals who are motivated to work on the project. In a distributed team, unless you can trust the team, it can never work. In Ekstep’s case, they work with engineers from big corporates to developers in small indie dev shops — all working towards the mission of educating under-privilidged children.</p><h4>Inclusion</h4><p>Trust can only be fostered if everyone’s goals are aligned. It also means that there is no distinction between partners and insiders — everyone after all is on the same team.</p><h4>Transparency</h4><p>The other way trust develops is through transparency. Though it might be argued that it is slightly easier for non-profits since there are no IPs to protect, EkStep however allows everyone to participate in design decisions and all metrics are available to everyone in the team. In general, if there is trust and tolerance for mistakes, transparency is easier to foster.</p><h4>Common set of processes</h4><p>EkStep engineering follows a common set of processes so that anyone can pitch in when required and there is no dependency on a single individual or a team. This includes things like tools being used, check-in processes, deployment to different cloud environments, and such.</p><h4>Extreme Automation</h4><p>Common processes also helps in automation, and automation further helps in defining more standard processes. As such there is no limit to automation — from automating infra and network set up, to writing test cases and CI. EkStep has done a fantastic job on this by modularising and containerising as much as possible. In an example they gave — they moved their complete infrastructure from AWS Singapore to Mumbai in 15 mins (though I am not sure how the data migration happened, and if that is automated as well).</p><h4>Focus on interfaces</h4><p>One interesting principle, clearly laid out was that during design discussions, the focus is more on the interfaces — defining the correct interfaces, contracts involved with the consumers of the interfaces and testing them. The actual implementation of interfaces is secondary and can be iterated over because the extent of its impact is limited.</p><h4>Scheduling</h4><p>With a distributed team, planning becomes very crucial. All the objectives and expectations are very clearly laid out in extreme detail at the start of a cycle and a release date is set upfront as well. The release date is non-negotiable and everything revolves around the date (which in EkStep’s case is set monthly). This helps the different stakeholders in the organisation plan their tasks.</p><h4>Always look for constant feedback</h4><p>EkStep looks for constant feedback at both an organisation level and an individual level. Whether it is getting feedback from AWS on their automation or security configuration, or a PM and dev looking for feedback on their progress, everyone is very open to feedback and are actively looking for it. One thing I liked was the idea of a weekly demo they conduct where all PMs and dev showcase their work, and anyone in the company can review it.</p><p>This discussion helped me think about other ways for building products and scaling engineering teams — an alternative framework to hiring full-time and/or outsourcing the less critical aspects of development.</p><p>Thanks to Pramod and Madhu from <a href="https://ekstep.org/">EkStep</a> for taking the time to have the conversation, and <a href="https://www.linkedin.com/in/theswamy/">Sanjay Swamy</a> and team from <a href="http://primevp.in/">Prime Ventures Partners</a> to organise the event.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2278e7b5f143" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Finance as a harbinger of data... and products (an engineer’s take)]]></title>
            <link>https://medium.com/clarity-in-chaos/finance-as-a-harbinger-of-data-and-products-an-engineers-take-c5233c488704?source=rss-44aef2fce762------2</link>
            <guid isPermaLink="false">https://medium.com/p/c5233c488704</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[finance]]></category>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[banking]]></category>
            <category><![CDATA[ideation]]></category>
            <dc:creator><![CDATA[Devang Mundhra]]></dc:creator>
            <pubDate>Thu, 15 Sep 2016 04:51:03 GMT</pubDate>
            <atom:updated>2016-09-15T04:51:03.345Z</atom:updated>
            <cc:license>http://creativecommons.org/licenses/by/4.0/</cc:license>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0xPsk28f3s6dViFdRzE2CQ.png" /><figcaption>From Finance to Data to Ideas to Products</figcaption></figure><p>Financial institutions are an important pillar for the growth of the economy. Individual, SMEs and large enterprises — all need access to capital to grow a step further from where they currently are, the services for which are provided by financial institutions.</p><p>But financial institutions have been there for ever, how does it concern an engineer or product person?</p><h4>Financial services —a running track, but also a spring board</h4><p>While providing financial services could be a good business in itself, and can be expanded to grow horizontally serving many different domains, it could also give an opportunity to someone, <em>to go deep and transform a sector.</em></p><p>Just like people are willing to share their personal data to get free access to social interaction, similarly people and companies are more than willing to share data to get easy and efficient access to capital.</p><p><strong>So entrepreneurs trying to get an <em>in</em> in a domain can start to use capital and financing as a way to get access to a market, collect data and create products using that data — thus creating a virtuous circle. It also gives them an available customer-base, and capability to influence behaviour about products using access to capital as a lever.</strong></p><h4>EXAMPLE 1 — OH THAT HURTS!</h4><p>Lets say you are trying to change how bone fractures are treated. Instead of the plaster of paris (or similar synthetic materials) casts that are messy, unhygienic and really itchy, you come up with a product for swanky 3D printed orthopaedic casts. Now you can go to market in two ways-</p><ol><li>Do some market research and build a product. Go to an orthopaedic doctor, demo the product, give him some commission, train the staff, and hope the doctor recommends and uses this for his/her patients. However, here there is lesser incentive for a doctor to change his status quo, and any commission would have to be borne by the patient at the end of the day, driving up the price of the product.</li><li>Tie up with an orthopaedic doctor to provide loans for medical procedures. The doctor is happy because he can get payments quickly and can up-sell his services, the patient is happy because he can manage his unplanned expenses. Collect lots of data around procedures, costs and requirements, and incorporate that into the product. Possibly provide better plans to doctors and patients for using this product.</li></ol><h4>EXAMPLE 2 — WHERE IS THE MILK?</h4><p>You want to bring about a change in the animal husbandry industry, but don’t know where exactly to start. Here are two scenarios to proceed-</p><ol><li>Talk to a few farmers and breeders and identify a problem. Create a product for them, get it tested with some early users and then find a channel to reach more people with the problem you have solved. Test your product-market fit.</li><li>Start with micro-financing for farmers and animal breeders. Collect lots of data around their practices, produce outputs, costs, raw materials, etc. Find issues in their processes that increases their uncertainty and cost. Design a product around it. Provide better financing rates to them for using your product.</li></ol><p>Health care, agriculture, education, electronics, automobiles, etc.. — similar philosophies can apply to many domains. And given a lot of these sectors are traditionally underserved categories, they would be very attractive for financial institutions (like banks and NBFCs) to partner as well.</p><h4>Data is the new soil</h4><p>Last decade has seen a monumental rise in data collected via various kinds of social interactions. A lot of products around personalization were built using that data. Data from finance promises to be much more potent to build interesting products — not only for individuals, but also for businesses and complete sectors.</p><h4>Flip side</h4><p>This access to data, however comes with its own set of problems. Building a business is not easy, and in finance there are additional challenges around regulations and wariness about novelty.</p><p>Also, for startups which are typically resource-strapped, does it make sense to diffuse their focus to build a financial service in addition to building the product. Is data enough of a reason to build a financial service?</p><p>In addition, with all the data collected as part of providing financial services, there is also an obligation to protect the privacy of the customer. Using the data in a manner that protects the privacy of a consumer while providing useful insights is a very active research area currently.</p><p>So for a startup looking to build a product, one way could be to start with providing financing options in that domain. This not only allows for capturing a lot of high quality data about the customer, which would be otherwise unavailable, but also lets the startup stay in constant contact with their customer base.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c5233c488704" width="1" height="1" alt=""><hr><p><a href="https://medium.com/clarity-in-chaos/finance-as-a-harbinger-of-data-and-products-an-engineers-take-c5233c488704">Finance as a harbinger of data... and products (an engineer’s take)</a> was originally published in <a href="https://medium.com/clarity-in-chaos">Clarity In Chaos</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Idea for cleaner air: Electric rickshaws in India]]></title>
            <link>https://medium.com/clarity-in-chaos/idea-for-cleaner-air-electric-rickshaws-in-india-48d539908e18?source=rss-44aef2fce762------2</link>
            <guid isPermaLink="false">https://medium.com/p/48d539908e18</guid>
            <category><![CDATA[climate-change]]></category>
            <category><![CDATA[electric-vehicles]]></category>
            <category><![CDATA[cleantech]]></category>
            <dc:creator><![CDATA[Devang Mundhra]]></dc:creator>
            <pubDate>Wed, 02 Dec 2015 11:05:50 GMT</pubDate>
            <atom:updated>2015-12-02T11:05:50.152Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*L-vVWl9Fn25RrVgXkemUZg.jpeg" /><figcaption>Auto rickshaws by <a href="https://www.flickr.com/photos/slyronit/">Ronit Bhattacharjee</a></figcaption></figure><p>In the backdrop of the Climate Change meet up going on at Paris, I thought why not share a thought that I have been brewing for some time about a way to alleviate this problem. Would love to get critical inputs and suggestions.</p><h4>The Problem</h4><p>In a previous blog, I had mentioned that bad air quality is a very tangible problem in India. <a href="https://en.wikipedia.org/wiki/Air_pollution_in_India">Though burning fuel wood and bio-mass is a big contributor to this, exhaust from vehicles comes a close second</a>. This blog is about that.</p><p>Electric vehicles is a dream dreamt by many. The job of a motor car is to convert any energy into mechanical energy in the most efficient way. In this case, it happens to be through battery powered motor. And the job of power plants is to convert any energy into electricity, so let them do that efficiently.</p><p>The biggest problem with electric vehicles is range anxiety, and that can be only solved through infrastructure that is not present in India yet. However, there is one class of vehicles that is aptly suited for this — auto rickshaws — which can kick start infrastructure development for battery operated vehicles.</p><h4>The Plan</h4><blockquote><em>Convert auto-rickshaws into electric, and provide battery as a subscription. Auto-rickshaw owners will pay for the cost of conversion upfront, and pay a monthly fees for charged batteries and maintenance.</em></blockquote><p><em>Notes:</em></p><ol><li><em>This battery subscription/swapping idea has been tried by quite a few players across the world (BetterPlace, Gogoro) with various degrees of success.</em></li><li><em>This is similar to how the government had asked rickshaw drivers to convert their vehicles into CNG. Unfortunately, </em><a href="http://papers.sae.org/952507/"><em>CNG is not all that clean and safe</em></a><em>, especially when the problem is about CO2 emissions in addition to pollutants.</em></li></ol><h4>Why Rickshaws?</h4><p>The characteristics of auto-rickshaws are-</p><ol><li><strong>Small to moderate range</strong>: They are mostly used for urban transport where the distances are not very large, after all you usually don’t go on a long drive in a rickshaw!</li><li><strong>Standard models</strong>: There are only a few manufacturers of auto rickshaws and their models are quite standard. So economics of scale comes in handy here.</li><li><strong>Clustering</strong>: Rickshaws usually cluster at certain locations, so providing swapping stations near those few locations can serve almost everyone.</li><li><strong>Maintenance</strong>: Rickshaws need regular maintenance, however, the cost of maintenance for them has been steadily increasing. Also, rickshaw manufacturers have a conflict-of-interest in helping keep maintenance costs low because spare parts are almost as big a business for them as sales. Electric vehicles, on the other hand, needs very low maintenance so that works in favor for them.</li><li><strong>Fuel availability</strong>: Rickshaw drivers relying on CNG need to queue up in long lines and waste time where they could be potentially earning money. This is because availability of CNG is centralized to a few pumps. Electricity, on the other hand is every where and so a lot less time will be wasted.</li></ol><p>With these characteristics and a few more, rickshaws seem to be ideal to start creating an electric fleet for urban transport.</p><h4>The Details</h4><p>To convert a petrol or CNG auto-rickshaw to electric, the following components are needed-</p><ol><li>DC <a href="https://en.wikipedia.org/wiki/Electric_motor">Motor</a> (~Rs. 30K)</li><li>DC-DC <a href="https://en.wikipedia.org/wiki/Motor_controller">Controller</a> (~Rs. 5K)</li><li><a href="https://en.wikipedia.org/wiki/Battery_management_system">Battery Management System</a> + <a href="https://en.wikipedia.org/wiki/Lithium-ion_battery">Li-Ion Batteries</a> + Battery Charger</li><li>Miscellaneous</li></ol><p>Of these, an auto-rickshaw driver will need to buy 1. + 2. <strong>Capital investment by rickshaw driver: Rs. 35K</strong></p><p><strong>Batteries</strong> are the capex for us and opex for the rickshaw drivers. Assuming a rickshaw plies ~120 kms per day on an average. If just a single battery pack is to be used in a single day, it would need approx. 15kWh battery pack.</p><p><a href="http://www.autoblog.com/2015/10/08/gm-li-ion-battery-cost-per-kwh-already-down-to-145/">Assuming $150/kWh</a>, this means <strong>each battery costs Rs. 150K</strong> (which is actually almost the cost of an auto-rickshaw).</p><p>So, how is this advantageous to the rickshaw driver? With the cost of CNG at ~Rs. 44/kg, and the mileage of a rickshaw at around 40kms/kg, plying 120 kms per day would cost Rs. 130 per day or Rs. 4000 per month. In addition, it costs ~Rs. 1000 per month for maintenance.</p><p><strong>Current monthly opex of rickshaw drivers: ~Rs. 5000</strong></p><p>On the other hand, <strong>with this approach, a rickshaw driver will have to pay less than Rs. 4000</strong> (assuming life of batteries is about 10 years, a 30% rate over a 10 year period will be ~Rs. 4000 monthly).</p><h4>The Challenges</h4><p>Some of the challenges in this venture are mentioned here, but all of them have either already been solved elsewhere, or can be solved in a scalable manner using technology.</p><ol><li><em>Security of batteries</em>. Batteries are the critical piece for this to work and are quite expensive. There is a chance that a few folks might be tempted to not return the batteries.</li><li><em>Reliable electricity supply.</em> An uninterrupted supply of electricity will be required to continuously charge the batteries, otherwise it could affect the livelihood of the rickshaw drivers relying on us.</li><li><em>Initial capex by rickshaw drivers. </em>Rs. 40K is a big initial investment to ask for from rickshaw drivers.</li></ol><h4>Final Words</h4><p>Electric rickshaws are already picking up steam in a few cities like Delhi. However, they use lead-acid batteries which have very poor range (less than 50 kms), very short life, unsafe especially at high temperatures, and would be a nightmare to dispose. For that, Li-ion or other solid-state batteries are better suited.</p><p>This is just one way to get the ball rolling on cleaner air in a way that is justified purely by numbers, without any altruism in the picture. Would love to hear your thoughts and ideas on this, or other ideas to fix the issue.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=48d539908e18" width="1" height="1" alt=""><hr><p><a href="https://medium.com/clarity-in-chaos/idea-for-cleaner-air-electric-rickshaws-in-india-48d539908e18">Idea for cleaner air: Electric rickshaws in India</a> was originally published in <a href="https://medium.com/clarity-in-chaos">Clarity In Chaos</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How it *really* feels to come back?]]></title>
            <link>https://medium.com/clarity-in-chaos/how-it-really-feels-to-come-back-28106ab0666c?source=rss-44aef2fce762------2</link>
            <guid isPermaLink="false">https://medium.com/p/28106ab0666c</guid>
            <category><![CDATA[moving-back]]></category>
            <category><![CDATA[problems]]></category>
            <category><![CDATA[india]]></category>
            <dc:creator><![CDATA[Devang Mundhra]]></dc:creator>
            <pubDate>Sat, 24 Oct 2015 06:48:36 GMT</pubDate>
            <atom:updated>2015-12-02T11:45:27.239Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*ZXCo2JX2b8jvn2gNOX2bEQ.gif" /></figure><p>My earlier post about coming back to India <a href="https://medium.com/clarity-in-chaos/how-it-feels-to-come-back-7a5d623bb28d#.hn08pk30m">had a bit of romanticism</a>. This is more for the realists — some rants and complaints. <em>EDIT: And some solutions I found in the past couple of days.</em></p><h4>Traffic is a bitch</h4><p>I could not drive after coming back. Atleast not immediately. There seems to be a secret unsaid language that everyone on the road speaks to figure out how to move on the road. I used to know this language, but have since forgotten.</p><h4>Follow-up is a pain</h4><p>Things break. They are built to be so. But what is not unreasonable to expect is to get it repaired quick and easy. But thats not so true in India. Every leaking tap and broken air-conditioner needs atleast a few phone calls per day to get the service person come and fix it. No wonder home services is one of the fastest growing category for startups — it desperately needs some organization.</p><h4>Air quality remains wanting</h4><p>The mornings are quite serene. Sun rising- its rays diffracted blood orange by the dust stirred up by street cleaners. Birds chirping- bicyclists ringing their bells. But suddenly a waft of smoke! Someone burning up leaves and litter leaving a trail of smoke. I don’t know how to solve this effectively, but I have a thought about how to improve air quality here. Will probably write about it sometime. <em>EDIT: </em><a href="https://medium.com/clarity-in-chaos/idea-for-cleaner-air-electric-rickshaws-in-india-48d539908e18#.h8oqtmrqn"><em>Here it is.</em></a></p><h4>Flavors and fragrance everywhere</h4><p>A walk on any street will easily put most of the senses in overdrive. The cuisines of the world, available on street side. Smell of water on dust, augmented by the fragrance of champa and mogra. The heat is oppressive but such small things brings oasis of respite.</p><h4>Customer service could use a brush up</h4><p>A lot of the places I went to with customer facing personnel (hello Vodafone, Customs), many people were busy shirking. Even if there was a queue of customers waiting to be served, they would avoid eye-contact with customers — busy doing nothing on the computer screen, and let the one person handle all the load. Only when a superior officer came, or an enterprising customer managed to get their attention, would they reluctantly begin to serve. It was quite apparent who were poised for growth at such places.</p><p>Ironically however, if a person managed to get you on the phone or via email, they just wouldn’t want to leave you!! Even repeated requests to unsubscribe doesn’t work (Hello Tata Indicom, HDFC!). <em>EDIT: No wonder Truecaller is one of the most famous apps here.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=28106ab0666c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/clarity-in-chaos/how-it-really-feels-to-come-back-28106ab0666c">How it *really* feels to come back?</a> was originally published in <a href="https://medium.com/clarity-in-chaos">Clarity In Chaos</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How it feels to come back?]]></title>
            <link>https://medium.com/clarity-in-chaos/how-it-feels-to-come-back-7a5d623bb28d?source=rss-44aef2fce762------2</link>
            <guid isPermaLink="false">https://medium.com/p/7a5d623bb28d</guid>
            <category><![CDATA[moving-back]]></category>
            <category><![CDATA[emotions]]></category>
            <dc:creator><![CDATA[Devang Mundhra]]></dc:creator>
            <pubDate>Fri, 16 Oct 2015 06:22:54 GMT</pubDate>
            <atom:updated>2015-12-02T11:57:10.297Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3SItncOWX6yzLGjkjDhzAw.jpeg" /><figcaption>Photo by <a href="https://www.flickr.com/photos/nishanthjois/">Nishanth Jois</a></figcaption></figure><p>One of the most common questions I get once I moved to India after more than seven years of living in the Bay Area is</p><blockquote>how it feels to come back?</blockquote><p>Feelings, I think, are notoriously difficult to describe. The simple ones like <em>happy</em>, <em>sad, in love </em>and such are easy, and indeed, there are simple 😊emoticons😔 to describe all these and more. These feelings are so common that everyone can relate to them.</p><p>But the more complex ones, are a bit more difficult. I think the best way to describe it would be to draw an analogy to something more common.</p><blockquote>Every now and then, you meet a friend whom you might not have seen for a long time. You could have been inseparable before, but somehow grew apart over the years following different paths. But when you meet him, you feel like it was just yesterday that you two were chilling out. The conversations seems to be a continuation of where you left off.</blockquote><p>Thats how I feel on my move back! :)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7a5d623bb28d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/clarity-in-chaos/how-it-feels-to-come-back-7a5d623bb28d">How it feels to come back?</a> was originally published in <a href="https://medium.com/clarity-in-chaos">Clarity In Chaos</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Chocolate, Strawberry or Vanilla?]]></title>
            <link>https://medium.com/clarity-in-chaos/chocolate-strawberry-or-vanilla-357868c7ec4b?source=rss-44aef2fce762------2</link>
            <guid isPermaLink="false">https://medium.com/p/357868c7ec4b</guid>
            <dc:creator><![CDATA[Devang Mundhra]]></dc:creator>
            <pubDate>Mon, 29 Jul 2013 20:27:52 GMT</pubDate>
            <atom:updated>2013-07-30T21:38:06.166Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*lDObMj8ZOnnTcBcM.jpeg" /><figcaption>Two kids inside icecream shop/Duyanh Pham</figcaption></figure><h4>On making choices</h4><p>Reading about technologies, products and services people create often amazes me, sometimes wishing <em>Oh man, I wish I was doing that! </em>Inevitably, my thoughts wander off to thinking what should I work on.</p><p>But then I stumble. How do I pick which problem to solve? Somehow, through some set of choices and some serendipity, I have reached here. What next?</p><p>Picking a direction is difficult. Oh-so-difficult. Which field to work in, what form should the product or service be manifested, what kind of customers to serve, etcetera. <a href="http://www.ted.com/talks/barry_schwartz_on_the_paradox_of_choice.html">With so many choices, it gets more and more difficult to commit to one thing</a>.</p><p>Some romantics tell me that it will be like love at first sight, you’ll know it when you know it. Some pragmatists tell me to just pick something — anything — and keep going with it.</p><p>I am hoping to find a framework to help make a choice when at the crossroads.</p><h3>Framing it right</h3><p>Most aspects of a business have frameworks to help them. Marketing — pick up the 4Ps, M&amp;A — get the Boston matrix, product development — would you like to be lean or agile?</p><p>So I wonder what dimensions can I arrange my thoughts on to figure out which opportunities to pursue.</p><blockquote>The optimization problem is an abstraction of the problem of making the best possible choice from a set of candidate choices.</blockquote><p>Mathematics never fails.</p><p>The first time I read this in a text book, I thought I had found the way to derive at the <a href="http://en.wikipedia.org/wiki/Answer_to_The_Ultimate_Question_of_Life,_the_Universe,_and_Everything#Answer_to_the_Ultimate_Question_of_Life.2C_the_Universe.2C_and_Everything_.2842.29">answer 42</a>. Turns out though, the skill involved is not so much how to solve the problem, but to form the question correctly.</p><p>An optimization problem has the form</p><p><em>maximize an </em><strong>objective function, </strong><em>subject to </em><strong>constraint functions.</strong></p><p>Goals and limitations. Nothing new. But the thing about these two dimensions is that they are closely related to each other. If there is some place you want to be, and you are not there, something is holding you back. What is that? Identifying the limiting factor is a step closer to the goal.</p><p>For every parameter in your objective, there always has to be a corresponding limitation. The exercise is in finding that tradeoff. If my goal is to build a good product and have enough time to meet with people and develop a hobby, what can I do? Maybe scrape some time from my swimming to meet more people, or be a little slower in building my product to take a woodcarving class.</p><h3>Applying the filters</h3><p>But hey, you say, not every thing can be solved algorithmically. True, how about the intangibles? Everyones got their gut feeling, their source of inspiration. And thats a filter worth applying to the choices.</p><p>Filters can come from anywhere. They can be your own. They can be inspired. Whats yours?</p><p>When I was in high school, my school gave each student a diary — for keeping notes about homework and class schedules, for teachers to send messages to parents, and other official things. As students, however, we used it for playing tic-tac-toes, dots and as a supply of paper to make planes and paper balls for guerrilla warfare…</p><p>Anyhoo, on the back of the diary were written these lines, also known as Gandhi’s talisman-</p><blockquote>“I will give you a talisman. Whenever you are in doubt, or when the self becomes too much with you, apply the following test. Recall the face of the poorest and the weakest man [woman] whom you may have seen, and ask yourself, if the step you contemplate is going to be of any use to him [her]. Will he [she] gain anything by it? Will it restore him [her] to a control over his [her] own life and destiny? &lt;snip&gt;<br>Then you will find your doubts and your self melt away.”</blockquote><p>I never paid much heed to them then, but having made a lot of planes during my classes, I read these lines over and over and they got stuck.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Krb-2T0TZC0DJqBE.gif" /></figure><h3>The final picture</h3><p>Start thinking about choices. Don’t wait for all the data to come in before making the decision, but take a step and course correct on the way.</p><p>How do <strong>you </strong>evaluate your options on what topic to work on, what company to start, what city to live in or what flavor gelato to take?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=357868c7ec4b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/clarity-in-chaos/chocolate-strawberry-or-vanilla-357868c7ec4b">Chocolate, Strawberry or Vanilla?</a> was originally published in <a href="https://medium.com/clarity-in-chaos">Clarity In Chaos</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>