Applied AI
Language systems that survive real product constraints
Model quality, inference efficiency, calibration, and maintainable experimentation treated as one engineering problem.
Explore sourceWork
I care about the full loop: framing the problem, building the model, measuring what matters, and making the result useful to people.
Applied AI
Model quality, inference efficiency, calibration, and maintainable experimentation treated as one engineering problem.
Explore sourceResearch
Published work exploring graph attention, selection, and generation for abstractive summarization.
Explore sourceFounder work
A product-building practice across AI systems, planning, human control, and trustworthy interaction.
Explore source