Neil Tripathi
CS from NYU (B.A., Minor in Math). I work on neural controllers built from real connectome data, runtime safety for robot learning policies, and temporal reasoning in language models.
About
I graduated from NYU with a B.A. in Computer Science and a Minor in Mathematics (Dean's List 2023-2025). My current work sits at the intersection of computational neuroscience and robotics: building locomotion controllers from real connectome wiring diagrams, testing whether biological circuits fail in more interpretable ways than learned policies, and deploying the results on physical hardware.
Alongside the neuro work, I build runtime safety tools for robot learning (TAP-Score), study temporal reasoning failures in frontier LLMs, and have two papers under review (VLM visibility benchmark with Prof. Ernest Davis at NYU, and radiotherapy robustness at SERA + ASTRO 2026). Previously I co-founded Physical.AI (top 6, YC W25 hackathon) and built Sympli, a voice-first health companion with a HIPAA-aware AWS stack.
Skills & tools
- AI/ML: PyTorch, MuJoCo, Brian2, diffusion models, LLM fine-tuning, object detection, evaluation pipelines, agent tool use
- Programming: Python, TypeScript, SQL, Java, Rust, C
- Data & Cloud: Postgres, MongoDB, BigQuery, Supabase, AWS, Docker, REST APIs
- Languages: English (fluent), Hindi (fluent), Spanish (intermediate), French (basic)