I'm a San Francisco-based founder, engineer, and researcher incubating at Harvard Innovation Labs. I work across artificial intelligence, machine learning, data infrastructure, full-stack engineering, product development, and robotics, with a particular interest in intelligent systems that must be reliable, measurable, and genuinely useful in the real world.
I build AI-native products end to end—from defining the research question and product strategy to designing data pipelines, model architecture, backend services, cloud infrastructure, and polished user experiences. My work includes LLM fine-tuning and serving, retrieval-augmented generation, adaptive agent behavior, model evaluation, safety and integrity controls, scalable APIs, observability, experimentation, and latency and capacity optimization.
As the founder of Classy AI, I designed and built a governed adaptive learning platform that combines Socratic tutoring, course-grounded retrieval, learner-state modeling, academic-integrity protections, and educator-facing evidence of understanding. Building the company has required me to work across research, systems architecture, product design, evaluation, deployment, pilot readiness, and company strategy.
My current robotics research explores closed-loop manipulation under uncertainty, combining RGB-D perception, symbolic planning, visual verification, motion control, and bounded learned recovery in a reproducible sim-to-real environment.
What I do best is turn ambitious, ambiguous ideas into rigorous systems that can be tested, deployed, and improved with real evidence. I'm especially interested in ethical, human-centered AI and in building technology that helps people learn, work, and create more effectively.
Always open to collaborating, learning something new,
or exploring interesting technical problems.