Nov 2021 – Nov 2022 · 1 yr

• PleDu is short for playful education. We are re-imagining STEM education for 21st century with 'Game Based Learning.' With PleDu, students play innovative, well-designed and fun video games and learn mathematics. And we are talking real video games (think Mario or Zelda or Shooter games), not just cute graphics.
• PleDu is currently in beta. You can try it out at https://pledu.co/ . 100s of kids (and adults!) have got much better intuition for topics like Trigonometry, Co-ordinate Geometry, Vectors. We've seen 1000s of minutes of gameplay on the platform already. In long run, we are building a Metaverse ready interactive education for learning STEM.
• PleDu can be used as a complement to classroom education and alternative to static books. Reach out via DM if you are interested to know more.

Sep 2018 – Jun 2021 · 2 yrs 9 mos · Pune/Pimpri-Chinchwad Area

• Worked as data engineer / data scientist on Slack’s Calls, files, calendar and email integration. Shipped multiple data pipelines that work at Slacks scale to production.
• Astro, the Palo Alto, California based startup I was working with earlier, was acquired by Slack in its biggest acquisition till date. I joined Slack as a part of the acquisition.

Mar 2018 – Sep 2018 · 6 mos

• Astro is making email and workplace communications more intelligent, so teams can focus on what matters. I work as machine learning engineer. My work involves writing, testing and maintaining machine learning algorithms which form the backbone of Astro's AI offering. Advances in artificial intelligence, machine learning, and natural language processing, enable Astro to bring improvements to email that weren’t possible even a few years ago. We believe AI will fundamentally change the way we use email.

Mar 2017 – Mar 2018 · 1 yr · Pune/Pimpri-Chinchwad Area

• Onlinesales.ai is SaaS business spun off from Sokrati for digital marketing in e-commerce. I am leading the data science practice for the company, building the core IP behind the product.
• Work profile involves conceptualizing new product and features powered by machine learning, owning the implementation-testing-validation of all machine learning algorithms. The team has built many algorithms which have made direct impact on top-line of many e-commerce companies in India and US. Few examples include,
• A large scale text classification algorithm using probabilistic models that allows us to derive robust industry benchmarks. The algorithm works on product catalogs with 4+ million products.
• Algorithm for budget management on advertising campaigns using Bayesian estimation and linear programming, which optimize 1000+ advertising campaigns daily.
• Improving search engine marketing of 80+ retailers using natural language processing and semantic similarity.
• Large scale recommendation engine for on-site advertisements using mixture of content based and collaborative filtering algorithms. This engine powers millions of ad impressions daily for large marketplaces in India and South-East Asia.
• I handled growing the team from individual contributor to 7+ data scientists & engineers. I also setup an innovative "new data scientist boot-camp" to ensure that data scientists with academic knowledge can work with engineering teams quickly and productively.

Oct 2012 – Mar 2017 · 4 yrs 5 mos · Pune/Pimpri-Chinchwad Area

• Sokrati, a leader in Paid Search & Social Advertising technology, provides a sophisticated solutions to efficiently manage digital marketing spends for a large number of businesses.
• The technology developed takes a beating of serving and tracking 20M+ impressions a
• day. It runs crunching data mining algorithms on tera-bytes of data using distributed
• computing, all developed in-house.
• My work at Sokrati is primarily focussed on being a master of multiple trades : statistics, machine learning, programming and data vizualization. The data team at Sokrati builds predictive models for forecasting, bidding and budget allocation for a large number of clients across domains like e-commerce, listings and banking. The algorithms manage advertising portfolio of millions of ads in completely automated fashion.
• I routinely work with data mining tools (R, Python are our favourite), standard as well as novel machine learning algorithms and techniques (random forests, neural networks, natural language processing, regression models, clustering, bayesian models) and big data technologies and NoSql data stores (Cascalog Hadoop/MapReduce, Apache spark, redis, MongoDB).

Aug 2011 – Oct 2012 · 1 yr 2 mos · Mumbai Metropolitan Region

• Part of analytics experts group overseeing customer value management activity for complete customer base of Vodafone India. I was involved in development, validation and management of plethora of machine learning models for revenue enhancement, churn prediction, data/value-added-services/internet adoption, customer segmentation, retailer segmentation etc.

Jul 2010 – Jul 2011 · 1 yr

• Analytics and consulting assignments for banking, insurance and financial services companies in India and oversees. Worked on variety of assignments like credit scoring models, cross sell/up sell models, lapsation and reinstatement models for insurance companies, customer segmentation, portfolio valuation and cash flow prediction.