About
Parth K. Thaker · AI Research Engineer
I'm an AI Research Engineer at Intuitive Surgical, where I build secure, privacy-preserving LLM systems — assistants over internal documents and images, guardrails and security analysis for LLM applications, and tooling that brings LLMs into CI/CD workflows.
Before industry, I did a Ph.D. at Arizona State University with Prof. Gautam Dasarathy, working at the intersection of nonconvex optimization, graph theory, and multi-armed bandits — from sample-complexity and landscape analysis for quadratic feasibility to graph-informed pure exploration (GRUB, NeurIPS 2022). I care about results that are both provable and usable, and I try to carry that discipline into the systems I ship today.
Focus
Experience
AI Research Engineer
Jul 2024 — PresentIntuitive SurgicalSunnyvale, CASecure, privacy-safe LLM workflows — document and image assistants, guardrail and security analysis for LLM applications, and LLM + CI/CD tooling — alongside platform and infrastructure work: pipeline performance, cross-region Kubernetes, and operational monitoring.
Ph.D. Researcher — Optimization & Bandits
Aug 2017 — Jun 2024Arizona State UniversityTempe, AZNonconvex optimization, multi-armed bandits, and graph theory with Prof. Gautam Dasarathy — from sample-complexity and landscape analysis for quadratic feasibility to graph-informed pure exploration (GRUB, NeurIPS 2022).
Research Intern
May — Aug 2023Intuitive SurgicalSan Francisco Bay AreaImproved a DINO-based self-supervised video-search system with fine-grained training objectives, raising retrieval quality under limited-label constraints.
Research Intern — Algorithms
May — Aug 2022Mitsubishi Electric Research Labs (MERL)Cambridge, MAIntersection of robotics and bandits for multi-agent resource monitoring under noisy sensing.
Systems Engineer
Sep 2016 — May 2017NetraDyneBengaluru, IndiaDriver-safety features from mobile sensor and IMU data to analyze driving behavior.
Summer Research Intern
May — Aug 2015Indian Institute of Science (IISc)Bangalore, IndiaLinear bandit algorithms under perturbation (validity of OFUL); a learning algorithm for multi-user bandwidth sharing, analyzed and proven to converge.
Summer Intern
May — Jul 2014Securifi Embedded SystemsHyderabad, IndiaSQL/NoSQL database performance analysis, cloud automation with Cloud-init + Puppet, and embedded-OS driver support on routers.
Engineering Intern
May — Jul 2013Cisco SystemsBangalore, IndiaA lightweight DHCP client (v4/v6) in Python to test BNG session controllers at 1M+ subscriber-session scale, with round-trip latency measurements.
Education
Ph.D., Electrical Engineering
May 2024Arizona State UniversityTempe, AZGraph theory and optimization, advised by Prof. Gautam Dasarathy.
B.Tech + M.Tech (Dual Degree), Electrical Engineering
2016IIT MadrasChennai, IndiaFive-year integrated dual degree; specialization in communication systems and signal processing.