Abacus is the world's first AI-assisted data science and MLOps platform and powers thousands of ML models in production in dozens of Fortune 500 companies.Our generative AI models will generate code in notebooks and program our end-to-end MLOps platform to launch AI systems in production.Our AI platform supports everything from enterprise-scale streaming, data-wrangling, and real-time feature store to neural architecture search, model monitoring, explainability, and drift.  Our domain-specific custom verticals help you create SOTA time series, recommender, NLP, predictive, and vision models.Key components of our platform include- enterprise-class machine and deep learning feature store- support for both supervised and unsupervised learning- neural architecture search for model building- ability to plug and play your own ML and DL models- enterprise-class model monitoring, bias detection, and explainability- notebook hosting and native python support Our research team regularly publishes at NeurIPs, ICML, and ICLR. We solve hard problems in AI, invent new techniques and program them into our state-of-the-art platform. Email me, if you want to move fast and invent things.

I started this division inside of AWS, with the intent of creating AI vertical services that apply AI to specific use-cases/domains. We conceived, designed and launched Amazon Forecast and Amazon Personalize. Forecast and Personalize use AutoML and deep-learning to create accurate custom deep-learning models for our customers. Both these services are first of their kind, and get us one step closer to truly automated machine learning and AI.In addition, we are working on a number of stealth projects in AI, Deep and Reinforcement Learning. My org. is looking for engineers, scientists and engineering managers who are interested in building large scale machine learning and deep learning systems.

(Acquired by Uber)Post Intelligence (Pi) is the world’s first deep-learning powered social media assistant. Pi helps users make great posts and gain followers by creating a deep learning model per social media user. Pi learns what you post about and what works well with your audience.Pi helps brands manage and optimize their social media. Influencers earn revenue by sharing relevant sponsored content suggested by Pi. More than 1.5m social media influencers used Pi and made over $100m in advertising revenue. As part of PI, we developed Candid, newsfeed app that used deep learning for moderation. The newsfeed app was used by more than 2m users. More about Candid - https://en.wikipedia.org/wiki/Candid_(app)Post Intelligence raised over $11m from 3 rounds of funding, and got acquired by Uber in 2017

Led the product division for a suite of Google Apps including Google Docs, Spreadsheets, Slides, Sites, Blogger, and Google Video. Created and launched Google Base (backend for Google Shopping) and introduced the name Google Shopping. Launched various quality-based bidding models and AdWords improvements as a Google Ads Product Manager.

Upwork was the fist services marketplace for the web. Was an early employee of Upwork and helped grow the company from a small start-up to $20M in annual revenue. Spent my first 2 years at Upwork as an software engineer and technical lead behind it's core marketplace infrastructure.

Technical lead for the Upwork/eLance marketplace. Technical lead for 20+ engineers on the first few versions of the Upwork marketplace

Developed computational models to identify and discover genes. Hold multiple patents for discovering cancer genes in the Drosophila melanogaster (fruitfly) and Caenorhabditis elegans (roundworms).