Yash Shetye

About Yash Shetye

Robotics Engineer with expertise in developing real-time perception systems for autonomous robots. Currently working on developing production-ready autonomous industrial systems combining robotics, computer vision and real-time control at BotCrew.Core expertise- Real-time computer vision & deep learning pipelines- Multi-sensor fusion & 3D reconstruction- C++/CUDA optimization for embedded systems- ROS2 & distributed robotics systemsPreviously worked on advancing distributed multi-robot perception systems at NYU Tandon\'s Agile Robotics and Perception Lab, integrating State-Of-The-Art Vision Transformers and optimizing CUDA/C++ implementations for embedded platforms. Also led computer vision initiatives at robotics startups, developing 3D SLAM solutions and autonomous navigation systems. Passionate about pushing the boundaries of computer vision technology for practical robotics applications.Open to opportunities in: Computer Vision Engineering | Robotics Software | Deep Learning

Experience

1. Robotics Engineer

   Botcrew

   Apr 2025 — Present

   Developing production-ready autonomous industrial systems combining robotics, computer vision and real-time control.Sensor Integration & Perception:• Porting Fast-LIO2 algorithm from ROS2 to standalone C++ for robust LiDAR-inertial odometry, providing stable pose estimation and point cloud output for proprietary navigation systems.• Developed custom C++ drivers for Hesai JT128/QT128 LiDAR and RealSense D457/D435i cameras with FastDDS middleware integration for real-time perception pipelines• Built point cloud processing pipeline (Open3D) with clustering and geometric analysis for autonomous I-beam-parallel navigation and path correctionML & Tracking Systems:• Deployed ROS2 C++ detection and tracking pipeline combining custom YOLOv8 models (99.5% mAP50) with Kalman Filter state estimation for robotic arm-mounted cameras autonomously tracking solar panelsIndustrial Robotics:• Programmed FANUC robotic arms (C++/Python) for autonomous sandblasting and painting applications, achieving 99.9% system reliability through optimized TCP/IP communication• Developed full-stack web applications (JavaScript/Vue.js) for robot control and real-time monitoringTechnologies: C++, Python, ROS1/ROS2, MoveIt, FANUC Robotics, YOLO, FastDDS, Open3D, JavaScript/Vue.js, Gazebo, RViz, Foxglove

Education

• NYU Tandon School of Engineering

  Masters, Mechatronics and Robotics

• Mukesh Patel school of technology management & engineering

  Bachelor of Technology , Mechatronics

  2015 — 2019