Build infra for LLM-based applications (Langchain, vector DB, and OpenAI/Llama2)

LLM model hosting and fine-tuning (e.g. Llama2)

Embedding model hosting and benchmarking (e.g. GTE)

Developed a feature store platform fostering cross-functional collaboration and efficient model training with tailored offline feature sets.

Engineered recommendation model serving infra for GoTo Suite, boosting personalized feature recommendations and enhancing customer retention.

Achieved Outstanding rating in two consecutive performance reviews for exceptional contributions.