- Worked on solving the Clinical Trail Matching problem for Cancer trails. Used NLP techniques to build a scalable and generalized model to rank the most relevant patients for each clinical trail (comparing criteria text with EHR records and Doctor's notes). Utilized open source Medical embedding weights (BioSent2Vec, etc) which are fed into an LSTM model to predict whether the input Criteria text and EHR records are a match.- Worked on De-noising of Seismic images using Conditional - Super Resolution - GANs, on a 6 month sprint. Deployed on Google Cloud Platform and developed both online and batch predictions for inference.- Worked on an exploratory project for Supply chain management using Reinforcement learning and Distributed systems. Experimented with multiple models including A3C (asynchronous-advantage-actor-critic), Deep Q-Network and DDPG (Deep Deterministic Policy Gradients)- Helped Quantiphi conduct two ML hackathon workshops for client executives, in collaboration with the Google Cloud team. Taught ML basics and helped participants work through a sample ML use-case in tensorflow. - Worked on an exploratory project for Anti Money Laundering and Anomaly/Fraud detection on simulated transactional data using Deep Auto-encoders.

Worked with Business Analysts to ensure in-time generation of weekly and monthly Business Intelligence reports that aid in business decisions. Used IBM Infosphere DataStage for ETL/data warehousing and Microstrategy for Business Intelligence Reporting & Analysis.