Probing the Safety of Large-Language Models (LLMs), and attacks on RAG enabled LLMs to identify state of the art threat vectors.

Building & deploying LLM-driven full stack applications across the organization. Developed tookits focused on NLP Topic Modeling for Chinese public equity research. Maintaining a internal Python package for LLM orchestration including CI/CD & unit testing.

Data visualization & analysis for the marketing team. Building & deploying ETL, data pipeline and Looker dashboard for internal stakeholders.