- Inferred text automation rules for medical reports by conducting data analysis on a large semi-structured biomedical dataset- Optimising a heart failure prediction model:Conducting time series analysis on cardiac implant data (stationarity & feature engineering)Optimising thresholds for categorical feature engineering through statistical methods- Enabled new KPIs in reporting tool by crafting specific tables from AWS datalake using Spark

- Implemented index to perform approximate nearest neighbor search on vector embeddings inside DBMS- Designed multiple tensor operations to enable complex pre/post-processing pipelines inside DBMS- Created multiple demos interacting with DBMS and custom ONNX runtime to remotely run complex ML models (YOLOv4, ResNext, GraphML)- Supporting GraphML models (API refactoring, improved error handling, exhaustive testing)

- Refactored multiple APIs to use a new service version while enabling retro-compatibility- Redesigned machine-dependent cron job for database update into cleaner API- Prototyping new market-data API using Stoplight Studio- Support operations team with analysis, automated scripts and testing

Teaching assistant for the Advanced Information, Computation, Communication 1 class taught by Prof. A. Lenstra