Leading a team to assess and predict the risk of on-road deployment using simulations.

Deepcell combines biology, hardware, mechanical engineering, software engineering and AI to build a product capable of identifying and isolating rare cells for ground breaking research and diagnostic. Built and led a high-performing software engineering team to develop a novel AI based cell sorting technology and established software engineering best practices and processes from scratch. Envisioned and led the development of the full software stack: • Data infrastructure on GCP (GCS, Airflow, Dataflow, Cloud functions) • Machine learning infrastructure for labeling, training and evaluation (Tensorflow, GPU/TPU, clustering, embeddings) • Instrument software with on device inference (Windows/C#, Linux/TensorRT, Python) • Web application (PostgreSQL, React, App Engine)

• Founded a start-up company with the vision of measurably changing user behavior towards higher retention and generating higher revenue using machine learning in the casual mobile gaming industry • Built a data ingestion, processing, ML training and prediction stack on AWS to become the backbone of a game personalization engine • Pivoted to building a scalable high-level Python and YAML-based system (https://github.com/productml/blurr) for machine learning feature engineering

• As part of the Google Ads team, led a team of 10+ engineers to build and launch an industry-leading online-to-offline store visit conversion product, making it possible for businesses to track how their ads translate to physical store visits. This highly impactful cross-functional project was directly responsible for generating an additional $XB annual revenue for Google • Conceptualized and evangelized the original idea of estimating store visit conversions based on observed data for Google Search Ads. Developed machine learning models and built pipelines and systems which operated on multi-petabytes of data with billions of data records • Built ad formats catering to mobile applications. Designed and implemented UI changes as well as backend integration with Google Ad systems

tenCube acquired by McAfee in Sep 2010. • Led the development of Android, BlackBerry and J2ME clients for WaveSecure, a mobile phone security product • Built a team to scale up the ability to make partner releases for Java based mobile platforms • Won 1st place in the Productivity/Tools category and 3rd place overall in Google Android Developer Challenge 2 (2010)

Encentuate acquired by IBM is Apr 2008. • Worked on the Windows client that provided Enterprise Single Sign-On by interacting with a remote server • Designed Windows Active Directory policies enforcement and multi-user sessions • Managed the Build System and the Installer • Provided pre-sales technical support followed by delivery and customization for enterprises