Current system
Benchmark and evaluation systems for VLM research
Built evaluation paths that keep model comparisons useful for curation decisions and model iteration.
San Francisco, CAMember of Technical Staff at Datology
Building multimodal data and evaluation systems at Datology.
I work across multimodal pretraining, VLM evaluation, distributed training, and research infrastructure. Recent projects span data curation systems, vLLM eval paths, multi-node training, and agentic tooling that shortens experimental loops.
The through-line is fast feedback from data decisions to trustworthy model comparisons.
Current work
Systems work on the critical path
Three systems lines currently carry the research loop from benchmark design through training and deployment.
Current system
Multimodal data curation and export pipelines
Built ingestion and export paths that make large multimodal corpora easier to train on and inspect.
Current system
Distributed training and launch infrastructure
Added vLLM eval support and hardened multi-node launch plus checkpoint behavior for faster experimental turnover.
Selected public work
Open work with the same systems bias
Two public artifacts that still reflect the same research-engineering direction as the current work.
