Daksh Thapar

About Daksh Thapar

PhD AI researcher building agentic AI systems for real-world businesses, with a foundation that goes back to pre-LLM deep learning. This lets me reason about LLMs and agents at a level most “prompt-only” practitioners cannot, and design solutions that are reliable, efficient, and actually deployable.Over the past years, I’ve worked across research and industry as:Staff Researcher at RadiusAI, leading large-scale computer vision initiatives and mentoring researchersLead Researcher at Fujitsu, delivering award‑winning 3D vision systems and multiple patentsResearch Scientist at Amazon, building production ML pipelines for customer-facing biometric productsPhD researcher at IIT Mandi, publishing 6+ papers in top-tier venues (CVPR, ICCV, ECCV, ACM-MM)What I’m great at:Turning messy, open‑ended business problems into well‑scoped AI systemsDesigning agentic architectures that combine LLMs, tools, retrieval (RAG), and classical MLBuilding robust computer vision and multimodal pipelines (video, 3D point clouds, biometrics)Bridging research and production: from idea → paper/prototype → deployed systemWhat I’m interested in now:Helping small and mid-sized businesses build their first serious agentic AI suite (support, ops, analytics, internal copilots)Designing and implementing end‑to‑end AI workflows, then training teams to use and maintain themHigh‑impact collaborations at the intersection of computer vision, LLMs, and real productsIf you’re exploring how agentic AI or advanced vision systems could fit into your product or operations, happy to brainstorm or review your ideas.

Experience

1. Staff Applied Scientist

   Radiusai

   Oct 2025 — Present · Bengaluru, IN

Education

• Indian Institute of Technology, Mandi

  Doctor of Philosophy - PhD, Computer Science

• Panjab University

  Bachelor of Engineering - BE, Computer Science

  2011 — 2015