Sahay Shivanker

I’m a Robotics and AI engineer focused on autonomous systems, intelligent control, and embedded machine learning—turning simulation work into hardware-ready capabilities. Right now, I’m building a simulation-first neurosurgical robotic planning workflow using ROS 2, MoveIt 2, Gazebo, Docker, and Python, including octant-wise reachability analysis and collision-safe motion strategies that mimic natural hand movement while respecting kinematic constraints.

I also develop end-to-end flight and control intelligence: I built an ML-driven offline Pixahawk configuration and optimization platform using Bayesian Optimization (Gaussian Process surrogate modeling) and PPO to automate PID gain search, cutting manual tuning from 3–4 hours to 20–30 minutes. On the embedded side, I’ve implemented STM32-based firmware in Embedded C with real-time Kalman-filter sensor fusion and deployed ML-optimized PID gains, achieving ±0.5° in-flight attitude deviation from setpoint under dynamic tests.