Deployed a full-stack AI webapp using Railway to provide personalized movie recommendations using Flask & React. I Developed a multi-threaded Selenium web scraper to efficiently scrape user rating history from their public Letterboxd profile. Using this data I trained an XGBoost model with 10+ features (Bag of Words/1 hot encoding) to predict user ratings on a 20000 movie dataset
Engineered a serverless batched-data pipeline with Python to ingest, normalize and analyze sports betting odds using Pandas to find +EV/arbitrage opportunities, with scheduled execution using AWS Lambda function.Setup real time alerts for profitable bets using Amazon SES and configured AWS CloudWatch to monitor performance.
Built a chrome extension to fact check tweets by parsing HTML and passing context into GPT 3.5’s API using ReactJS. Used prompt engineering to get specific answers in appropriate formats from GPT. Developed a simple UI to display the output with the additional option to view the full output.
