Constructed ground truth datasets to evaluate hybrid Retrieval-Augmented Generation (RAG) strategies, enhancing the accuracy and reliability of AI-powered information retrieval systems. Additionally, I implemented and tested various hybrid RAG approaches, optimized performance through exploratory code, and composed comprehensive technical documentation detailing methodologies, findings, and insights.

Developed an API using Python to execute SQL queries from natural language inputs, enhancing user accessibility to database queries. Additionally, I integrated OpenAI large language models with prompt engineering techniques to ensure accurate SQL query outputs and designed endpoints for secure and efficient file upload and query execution.

Learned from industry professionals and gained hands-on experience in a Y-Combinator funded cybersecurity startup, acquiring insights into innovative authentication solutions and day-to-day software startup operations. I designed and implemented the company’s website using HTML, CSS, and JavaScript, ensured a responsive and user-friendly experience, and assisted in creating and managing mailing lists and newsletters, contributing to the company’s outreach and communication strategy.

Gained hands-on experience in business disciplines through projects and presentations under the guidance of Berkeley Haas professors and corporate guest speakers during the Berkeley HAAS Business Academy for Youth Summer Program. As the CTO for my team’s startup idea, Concord, a mobile app promoting mental health, I developed and presented a comprehensive business plan, achieving 3rd place overall, while enhancing communication and teamwork skills by collaborating with international peers on market research and computer lab assignments.