Tech Lead, Coding Assistance

• Domain-specific fine-tuning to adapt LLMs for developer productivity

• Metric-driven end-to-end optimization to improve reliability and adoption of AI-assisted coding

Founding member of the Applied AI Team

• Team behind the creation of Databricks Dolly

• Early work on adapting LLMs to Databricks product applications

• AI-driven infrastructure optimization for Databricks serverless products

TLM of Marketplace Matching Inference: Team optimizes the rider/driver matching decision through Machine Learning, Algorithm Solving, and Scalable Backend Design.

Mattermark is using machine learning and big data techniques to collect, organize and analyze world's business information.

# Applied NLP to extract start-up funding data from raw news articles on the web. Used techniques include deep neural word embedding (Word2Vec) and classification model.

# Implemented feature extraction service that converts company information to input vector for ML pipelines.

# Built automated ML system to tag the industries of 4 million+ companies in Mattermark's database.

# Created general-purpose ML pipeline that trains company classification model and generates prediction automatically.

Translated the book Pattern Recognition and Machine Learning(by Christopher Bishop) to Korean