Summary
Harsh Munshi is an AI software engineer based in Munich with 10 years of deep experience in computer vision, robotics, and machine learning, blending research rigor with product-facing delivery. He has built end-to-end agentic LLM systems, custom tool integrations, and knowledge-graph-backed agents while improving prompt intent recognition accuracy by 80%. His background includes shipping real-time CV solutions—from 3D point-cloud pipelines and embedded face detection to client-side WebGL keypoint inference that boosted iPhone throughput 5x. He founded ResearchAgent.de and drives production ML at eigenblue, where he architected token-streaming websockets and automated LLM benchmarking and data collection. Comfortable across cloud stacks, edge devices, and frontend rendering, he pairs hands-on engineering with research-driven innovation. A lesser-known strength is his track record of turning novel research ideas (like OffsetNet) into robust, high-performance product components.
10 years of coding experience
7 years of employment as a software developer
Master of Science - MS Autonomous Systems, Master of Science - MS Autonomous Systems at Bonn-Rhein-Sieg University of Applied Sciences
AISSCE Science, AISSCE Science at Maharaja Agrasen Vidhyalaya
Master of Science (MSc) Computer Control and Automation, Master of Science (MSc) Computer Control and Automation at Nanyang Technological University Singapore
Bachelor of Technology (BTech) Instrumentation and control, Bachelor of Technology (BTech) Instrumentation and control at Nirma University
J H Ambani
English, Hindi, German