Jeff Mahler
Co-Founder, CTO at Ambi Robotics
Berkeley, California, United States
Summary
👤
Senior🎓
Top SchoolJeff Mahler is a robotics and computer vision leader with 12 years of experience building AI-powered manipulation systems and shipping them from lab to market as Co-Founder and CTO of Ambi Robotics. He combines a PhD-level research background from UC Berkeley in robotic grasping and sim-to-real transfer with hands-on engineering—contributing core GQ-CNN modules used for grasp candidate generation and ROS-integrated deployment. Jeff has repeatedly bridged academia and product: postdoctoral benchmarking of Dex-Net informed the commercial grasping stack he now architects. His early work in real-time 3D scanning and embedded pipelines (C++/CUDA) underpins a pragmatic approach to constrained compute on robots. Based in Berkeley, he excels at turning large-scale simulation, geometry, and ML research into robust production systems for warehousing and assistive applications. An understated strength is his track record of founding multiple ventures that translate cutting-edge perception research into deployable robotics products.
12 years of coding experience
9 years of employment as a software developer
Doctor of Philosophy (PhD), Electrical Engineering and Computer Science, Doctor of Philosophy (PhD), Electrical Engineering and Computer Science at University of California, BerkeleyElectrical Engineering, Computer Engineering, Electrical Engineering, Computer Engineering at The University of Texas at Austin
Github Skills (7)
machine-learning10
python10
machine-learning-models10
tensorflow9
computer-vision9
object-detection8
ros8
Programming languages (3)
Github contributions (5)
BerkeleyAutomation/gqcnn
Apr 2017 - May 2019
Python module for GQ-CNN training and deployment with ROS integration.
Role in this project:
Back-end & ML Engineer
Back-end & ML EngineerContributions:511 commits, 29 PRs, 325 pushes in 2 years 1 month
Contributions summary:Jeff's commits focus on developing core classes and modules related to the GQ-CNN, a Python module for training and deployment with ROS integration. They have implemented the foundational ImageGraspSampler class, crucial for generating grasp candidates. They also introduced a base Policy class and a specific Antipodal Grasping Policy, demonstrating an understanding of both algorithm design and application within a machine learning framework. The user contributed key code defining the fundamental components, like the FC layers of the GQ-CNN.
roboticspythondeep-learningdeploymentmachine-learning
jeffmahler/GPIS
Apr 2014 - Jun 2017
Contributions:453 commits, 1 PR, 297 pushes in 3 years 2 months
implicitimplicit-surfacesgaussiansurfacesgaussian-process
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.