I build the infrastructure behind frontier multimodal AI.
Software & ML systems engineer and Ph.D. scientist. I build production systems end-to-end — full-stack services, cloud infrastructure, data pipelines, evaluation, and agentic systems — with a focus on frontier multimodal AI. The demos here are original, built from scratch to prove it.
Full-stack, end-to-end, frontier-grade.
Multimodal
Vision/OCR understanding, text-to-video & speech generation, document and signal analysis.
Data curation & quality
Acquisition, filtering, dedup, human + synthetic curation, and quality studies at scale.
Evaluation & reward
Rubric systems, LLM-judge loops, benchmarks, ELO / win-rate / slop analysis, calibration.
Agentic systems
Tool-use loops, planner/worker/judge orchestration, self-correction, real-time streaming.
Full-stack & cloud
Production services & APIs (FastAPI, async, real-time streaming), cloud fleets (GCP/Docker), high-throughput concurrency, zero-downtime deploys.
Scientific depth
Ph.D. + peer-reviewed publications; rigorous study & evaluation design across STEM domains.