• Designed and optimized machine learning models (CNN/LSTM) to predict the emotions of images and songs using

PyTorch with hyperparameter tuning and transfer learning, and deployed them for real-time inference on web app

• Prototyped indicator-based trading algorithm with RSI, Bollinger Bands, and moving averages on QuantConnect, and

backtested across multiple time periods with over 1.6 Sharpe Ratio and 100%+ return

• Experimented and trained Reinforcement Learning agents in Forex trading achieving consistent 10%+ returns

• Designed and developed personalized benchmark HTTP API for analytics charts on Creator Dashboard using

microservices architecture with Nomad, gRPC protocols, and Amazon SNS/SQS with ASP.NET

• Migrated entire Creator Dashbaord React frontend to new UI library, improving consistency and render performance

supporting 9+ million creators with 8000+ DAUs on migrated pages

• Coordinated user acquisition tracking across teams, and authored tracking features in web and cross-platform client

that attributed source data to 200+ million hours of total playtime per month

• Built highly flexible GraphQL API backend with Django for course catalog with 100,000+ unique users annually

• Led and developed customizable user profiles, JWT Auth with Google OAuth, and course scheduling system to provide personalized student experience and features such as saving classes and creating custom schedules

• Automated data scraping of 2000+ courses with real-time enrollment updates and email notifications