Mayur Samrutwar

I’m a founder and software engineer focused on shipping end-to-end AI products—where simulation, learning systems, and user-facing tooling come together. At Isaac Protocol, I built a browser-based robotics IDE integrating MuJoCo physics with real-time 3D visualization, so developers can design and test robotic behaviors quickly.

I’ve also built learning and automation pipelines that move from research ideas to working systems. I developed a reinforcement learning pipeline using PPO for multi-task robotic policies, strengthened with imitation learning and domain randomization, and I implemented structured simulation environments with IK-based control and clear action interfaces. To make robot creation practical, I designed a visual editor that generates MJCF robot models from structured 3D components.

Earlier, I founded Hyre to deliver an AI interview experience end-to-end: LLM dialogue plus speech synthesis, a browser-based recording and processing pipeline using FFMPEG and AWS S3, and an automated candidate evaluation workflow driven by transcript parsing and structured scoring. I also strengthened my engineering rigor through security and testing work at Hilti Technology Solutions, where I performed penetration testing with Burp Suite, automated vulnerability reporting (cutting report generation time from 1 hour to 1 minute), mentored a TypeScript intern, and implemented integration tests with Robot Framework and Zephyr Scale.