Designed and implemented a chatbot development platform from scratch, spanning frontend (HTML/CSS/JS), backend (Python/REST APIs), and advanced subsystems (retrieval-augmented generation, knowledge graph–based memory).
Developed abstract model integration interface in Python (via Pydantic), enabling seamless model upload with minimal boilerplate; supported custom RAG pipelines with optional retrieval methods.
Built and deployed secure user management system with role-based access controls, JSON Web Tokens, and group based permissions tailored for healthcare data security.
Collaborated with data scientists and ML engineers to ensure usability.
Explored and implemented knowledge graph–based memory management to reduce conversational context bloat, outperforming conventional off-the-shelf solutions.
Produced clear documentation and interfaces emphasizing usability and maintainability.
Applied machine learning and NLP knowledge from graduate coursework (e.g., LLMs, chatbot design) to solve open-ended technical challenges.
Researched and implemented advanced ML Models, largely combinations of
CNN, LSTM, Transformer, GNN, and XGBoost
Collected and processed Satellite data to be used in advanced ML algorithms, predicting floods, deforestation, atmospheric conditions, and more
Created front and back end of an application to allow the user to select area of interest, and model to be used, with automatic download and processing of large amounts of satellite data