- Research Intern
- National Institute of Advanced Industrial Science and Technology (AIST)
- May 2026 - Present
- Tokyo, Japan
- Designing an agentic benchmarking platform in which each run executes inside a Docker sandbox, letting the agent provision its own test environment and run benchmarks there; targets multi-step reasoning and tool-use tasks
- Software Engineer
- InfoIMAGE, Inc.
- Mar 2024 - Apr 2025
- Brisbane, California
- Built an internal Core platform with Next.js and Flask, making data sample lookup 80% faster
- Designed a SQLAlchemy ORM and RESTful APIs, improving data consistency by 50%
- Automated a Python data-parsing pipeline, cutting manual mapping by 50%
- Software Engineer
- Tricia 42 LLC
- May 2024 - Nov 2024
- Remote
- Led 3 developers building the Next.js chat interface for the Tricia platform
- Implemented RAG with Supabase and a vector database, using the Vercel AI SDK for chat
- Machine Learning Engineer
- Amazon Alexa Prize / University of California, Santa Cruz
- Sep 2022 - Sep 2023
- Santa Cruz, California
- Fine-tuned RedPajama 3B with LoRA on AWS SageMaker, cutting training time 23% and raising user ratings 30%
- Deployed an auto-scaling inference server on AWS EC2, making the system 34% faster
- Improved user satisfaction 13.5% through A/B testing of dialog managers
- Data Scientist Intern
- HPC-FAIR / Lawrence Livermore National Laboratory
- Jun 2022 - Sep 2022
- Livermore, California
- Fine-tuned CodeBERT on DataRaceBench for parallel C/C++ code, with 10% faster training by freezing layers
- Built a Scrapy pipeline that collected data 25% faster for SARS-CoV-2 inhibitor analysis
