Robotics · Physical AI · Full-Stack Cloud
I build software where AI, robotics, and the physical world meet.
I'm Abraham Dauhajre — a founder and engineer with 15+ years building robotics, simulation, cloud, and AI/ML systems. I led robotics simulation at Amazon Lab126, the global content-metadata pipeline behind every title on Amazon Prime Video, and engineering at Formant and Savioke. I'm currently building a stealth startup in embodied AI with hybrid robotics/cloud infrastructure.
Inventor on 6 U.S. patents in haptics & sensing. Co-author of a 2026 embodied-AI paper on realistic synthetic household data generation. B.S. Electrical Engineering, University of Pennsylvania (control systems & robotics).
Built & shipped at
I've spent 15+ years at the boundary of software and the physical world — the domain now called physical AI — building robotics, simulation, cloud, and AI/ML systems.
I led robotics simulation at Amazon Lab126, led the global content-metadata pipeline behind every title on Amazon Prime Video, and served as Head of Engineering at Formant, a cloud robotics startup, where I built cloud robotics infrastructure and engineering teams from the ground up. Earlier, I was Director of Cloud Infrastructure at Savioke and worked on consumer devices and AR at CastAR, Immersion, and BlackBerry.
My work spans the full stack of getting technology from prototype to the real world: robotics, simulation, computer vision, cloud and data infrastructure, deployment, and observability.
Field Notes
I write and talk about building robots for the real world.
Essays, talks, and research on cloud robotics, sim-to-real, fleet operations, security, and what actually breaks between demo, pilot, and production.
-
Essay
Security by Design: End-to-End Encryption and Identity for Robotics Systems
Why perimeter/VPN security fails for robot fleets, and how to do end-to-end encryption and identity instead: provisioning tokens, per-call JWT identity, mutual TLS, and gRPC encryption on the LAN.
-
Paper
Realistic Synthetic Household Data Generation at Scale
Co-authored research on generating realistic synthetic household data at scale: training and evaluation data for embodied-AI and robotics simulation.