SQUARE Link to heading

SQUARE As the Technical Lead for SQUARE, an Enterprise Question Answering System, I spearheaded the development of a sophisticated question answering system utilizing extractive QA and retrieval augmented generation (RAG) techniques. SQUARE is designed to automate the extraction of information from extensive legal and medical documents using advanced knowledge graphs. My contributions include:
  • Developing a library to automate information extraction, significantly enhancing efficiency and accuracy.
  • Designing, implementing, deploying, and maintaining multi-tenant, serverless AWS environments to ensure scalability and reliability.
  • Creating an evaluation and prompt-tuning framework to optimize the performance of large language models (LLMs), further improving the system’s accuracy and user experience.

SQUARE is an ongoing project that continues to evolve and adapt to meet the needs of our enterprise clients.

MoleculeAI Link to heading

MoleculeAI In my role as Lead Data Scientist for MoleculeAI, an AI Drug Discovery Platform, I led the development of novel graph neural network models to support the discovery of medicinal psilocybin derivatives. Key achievements include:
  • Developing a molecular property prediction repository to automate model tuning and evaluation, significantly speeding up the drug discovery process.
  • MoleculeAI is currently seeking acquisition, with the platform already demonstrating its potential to revolutionize the early stages of drug discovery.

Split.it Link to heading

Split.it Split.it is a ridesharing platform I co-founded, designed to facilitate cost-effective travel options for users. As a Co-Founder, I led a development team of 7, achieving significant milestones:
  • Garnering over 250 active users and facilitating more than 400 rideshare trips.
  • Developing a user-friendly web interface, accessible across desktop and mobile platforms as a Progressive Web App, ensuring a seamless user experience.

This project was active from January 2022 to April 2024.

CourseChat Link to heading

CourseChat, an Independent Study project, is a Chatbot Assistant designed to help students with coding questions. As a Data Scientist on this project from January 2024 to May 2024, I was responsible for:

  • Creating a vector database and implementing an RAG orchestration layer to enhance chatbot performance.
  • Developing an automated evaluation pipeline for prompt engineering using user feedback, significantly improving the chatbot’s ability to assist students effectively.

Product Placement Link to heading

As a Data Scientist for Product Placement, a Proof-of-Concept project, I innovated by creating a tool that leverages Stable Diffusion to automatically generate realistic shadows and reflections, enhancing visual content. My contributions include:

  • Designing a serverless AWS architecture to deploy low-cost inference APIs, demonstrating the project’s feasibility and potential for scalability.

This project showcases my commitment to pushing the boundaries of AI and machine learning in practical applications.