Robot Learning Intern (Long-Term) İş İlanı

Robot Learning Intern (Long-Term) The Position We are looking for a motivated, hands-on intern to help us push the boundaries of what our autonomous robots can do. You will work on improving how our robots perceive the world and manipulate objects using a single robotic arm. Your code will run on real hardware where speed and precision matter — with mentorship from experienced engineers, but with real ownership of your tasks. Open to students currently pursuing a BS or MS or new graduates. What You'll Do Implement single-arm manipulation : Build and test algorithms for pick/place, insertion, and tool-use tasks on our robot arm, and help make them perform reliably beyond lab conditions. Support data collection and model training : Run tele-op sessions, curate datasets, and help fine-tune manipulation policies for robustness to real-world variation (object pose, lighting, occlusion) and toward tight tolerances and cycle-time targets. Build experiment tooling : Calibration, logging, evaluation, and visualization scripts that speed up iteration loop. Must Haves Educational Background: Currently pursuing a BS or MS in Robotics, Computer Science, Electrical/Mechanical Engineering, or a related field. ML Foundations: Solid understanding of machine learning fundamentals and modern learning-based manipulation methods — VLA models, Diffusion Policy, flow matching (coursework, research, or projects; prior deployment on a physical robot not required). Coding: Solid Python; comfortable working on Ubuntu/Linux. ML Frameworks: Good knowledge of PyTorch. Bonus Points Experience with ROS/ROS 2 (e.g., MoveIt). Hands-on robotics project experience — real hardware or simulation (labs, competitions, personal projects). RL or classical motion planning/control experience. Familiarity with computer vision and pose estimation frameworks. Hands-on "hacker" experience: you aren't afraid to debug a loose wire or calibrate a camera yourself. Research experience, publications, or published code we can look at (GitHub).