Summary
Tarik Kelestemur is an Applied Scientist at the Boston Dynamics AI Institute with 11 years of experience building vision-driven, whole-body mobile manipulation systems for real-world robotics. He holds a PhD in Computer Engineering from Northeastern University where his research fused probabilistic methods, implicit neural representations, and tactile sensing to enable sample-efficient visual-tactile pose and shape estimation used in precise manipulation. Before BD-AI, he was a Machine Learning Engineer at Covariant, troubleshooting and improving perception and grasping stacks across deployed pick-and-place robots and bridging simulation-to-reality data pipelines. Tarik has a strong full-stack robotics background—from TrajOpt trajectory planners and Gaussian processes to deep RL and Mask-RCNN—plus hands-on experience creating data generation, augmentation, and deployment workflows. He co-hosts an AI/ML podcast, signaling a knack for communicating complex ideas, and has repeatedly combined classical modeling with data-driven learning to tackle contact-rich robotic tasks.
11 years of coding experience
7 years of employment as a software developer
Bachelor of Engineering (B.S.), Electrical and Electronics Engineering, Bachelor of Engineering (B.S.), Electrical and Electronics Engineering at TOBB Ekonomi ve Teknoloji Üniversitesi
Doctor of Philosophy - PhD, Computer Engineering, Doctor of Philosophy - PhD, Computer Engineering at Northeastern University
Turkish, English