Founding Engineer: Contact Physics, Sim-to-Real & RL

Pixoft — Founding Engineer: Contact Physics, Sim-to-Real & RL (Europe / Remote or Hybrid)

What we are building:

Pixoft is building the touch and force-control foundation layer for industrial robots and humanoids. Robots can already see and move, but contact-rich manipulation (insertions, press-fits, connector mating, deformable/cluttered handling) still breaks vision-only models and requires weeks of manual tuning. Because force data cannot be scraped or recorded at scale, we generate synthetic contact data directly from CAD in GPU physics simulation, train compact RL policies with a contact curriculum, and deploy ~2MB skills directly onto existing robot controllers (KUKA, ABB, FANUC, UR) with a rapid on-cell real-to-sim calibration loop. We are backed by an a16z scout fund, Alchemist Chicago, and top industrial operators, with active pilots and partnerships across KUKA, Henkel, and Amazon Robotics.

Who we are looking for (1–2 Founding Engineers across Europe):

We are looking for a hands-on builder and researcher at the intersection of **Physics Simulation, Synthetic Data, and Reinforcement Learning for Manipulation** who wants founding-team ownership:

GPU Physics & Contact Simulation:

Deep hands-on experience with NVIDIA Isaac Lab / Isaac Sim, Newton, MuJoCo, or Brax/Warp—specifically modeling contact-rich dynamics, friction, tolerances, and synthetic force/tactile sensor generation from CAD.

RL for Manipulation & Sim-to-Real:

Proven track record training policies that actually cross the sim-to-real gap onto physical robot arms/hands (domain randomization, system identification / real-to-sim calibration, curriculum learning, and model distillation into compact edge policies).

Robotics Control Fundamentals:

Strong grasp of impedance/admittance control, force-torque feedback loops, and real-time execution on physical hardware.

Background:

PhD or Senior/Staff Engineer from top European robotics labs (e.g., ETH Zurich RSL, TU Delft, TUM, KIT, DFKI, INRIA, Imperial) or physical AI / robotics companies (e.g., Intrinsic, NVIDIA, Sereact, Agile Robots, Wandelbots, NEURA, DeepMind).

Why join now:

Founding-team equity, competitive compensation, immediate access to multi-vendor industrial robot testbeds (KUKA, UR, ABB) and European supercomputing clusters, and the chance to co-architect the core force foundation model (FM-1) from day one.