OpenAI· Robotics· San Francisco
Simulation Realism Engineer
Comp$293K – $385K
Classified Tasks (22)
Automate 0%Augment 82%Human-Only 18%
Augment (18)
AI assists, human decides
Define and operationalize realism metrics and validation protocols
analytical
Design experiments and automated tests to detect specific areas of non-realism
technical
Quantify realism gaps and track regressions over time using defined metrics
analytical
Fine-tune physics engine parameters (contacts, friction, mass/density, solver settings) to match measured real-world dynamics
technical
Adjust and optimize object and asset models so simulated dynamics match measured reality
technical
Evaluate third-party physics, rendering, and sensor simulation engines and benchmark their features
analytical
Integrate third-party simulation engines into the toolchain and extend them when necessary
technical
Create standards and guidelines for authoring and validating asset physical and visual properties
operational
Develop interactive tools for authoring and validating asset properties
technical
Implement semi-automated or fully-automated pipelines to tune and optimize assets for target engines
technical
Troubleshoot and resolve engine runtime issues in cloud environments, including OS, drivers, and GPU problems
technical
Parallelize simulation workloads by batching runs per engine instance to increase throughput
technical
Harden real-world-facing pipelines to ensure simulations run reliably at scale
operational
Compute, report, and analyze realism metrics as part of validation and regression testing
analytical
Partner with research teams to correlate simulation improvements with downstream model performance
analytical
Document and scale best practices for simulation realism across teams
communication
Produce ergonomic templates, documentation, and standards to enable engineers and researchers to author realistic tasks and assets quickly
communication
Operate simulation infrastructure at scale, including cloud ops, CI/HIL pipelines, and distributed GPU workloads
operational
Human-Only (4)
Requires human judgment
Own strategy, tooling, and operational practice to close simulation-to-reality gaps across physics, sensors, and rendering
leadership
Collaborate with research, software, hardware, and operations teams to identify highest-impact realism gaps
communication
Lead vendor proofs-of-concept and benchmark demonstrations to influence vendor roadmaps
leadership
Own and execute validation campaigns including teleoperation tests and hardware-in-the-loop (HIL) regression suites
operational
Job description
Simulation Realism Engineer | OpenAI Careers ## Simulation Realism Engineer Robotics - San Francisco Apply now(opens in a new window) **About the Team** Our Robotics team is focused on unlocking general-purpose robotics and pushing towards AGI-level intelligence in dynamic, real-world settings. Working across the entire model stack, we integrate cutting-edge hardware and software to explore a broad range of robotic form factors. We strive to seamlessly blend high-level AI capabilities with the constraints of physical systems to improve peoples’ lives. **About the Role** We are hiring a Simulation Realism Engineer to drive our effort to make simulation quantitatively real for robotics research and product development. This role owns the strategy, tooling and operational practice that close sim→real gaps across physics, sensors and rendering. You will work cross-functionally with research, software, hardware and ops to discover the highest-impact realism gaps, define measurable realism metrics, and deliver engineering solutions — from tuning engine parameters and asset standards to integrating third-party engines and operating sims at scale. This is a hands-on engineering role that blends scientific rigor (measurement & validation) with pragmatic systems work (tooling, cloud ops, and CI/HIL pipelines). **This role is based in San Francisco, CA. This role will require 4 days in the office per week and offer relocation assistance to new employees.** **In this role, you will:** * Define and operationalize realism metrics & protocols. Design experiments and automated tests to find specific areas of non-realism, quantify gaps, and track regressions over time. * Fine-tune engine parameters (contacts, friction, mass/density, solver settings) and object models so simulated dynamics match measured reality. * Evaluate, integrate, and — where necessary — extend 3rd-party physics, rendering, and sensor simulation engines (e.g., Isaac, PhysX, MuJoCo, video renderers or sensor sim frameworks). Lead vendor POCs and benchmark features to influence roadmaps. * Create standards, guidelines, and interactive tools for authoring and validating asset physical/visual properties; implement semi- or fully-automated pipelines to tune and optimize assets for target engines. * Solve technical issues for running engines in the cloud (OS, drivers, GPU), parallelize simulations (batching many runs per engine instance), and harden real-world-facing pipelines so sims can run reliably at scale. * Own validation campaigns (teleop, HIL regression suites, metric computation) and partner with research to close the loop between simulation improvements and downstream model performance. * Document and scale best practices. Produce ergonomic templates, docs, and standards so dozens of engineers and researchers can author realistic tasks and assets quickly. **You might thrive in this role if you:** * Have a strong mathematical & physics background — rigid/soft-body dynamics, contact mechanics, or FEA — and are comfortable reasoning about solvers and numerical stability. * Have hands-on experience with at least one or two production physics engines and sensor/render stacks (MuJoCo, PhysX, Isaac Sim, Newton, GPU renderers, or equivalent), and know how to measure/interpret their failure modes. * Enjoy plumbing every layer of the sim stack: CAD/asset import, collision/visual mesh tradeoffs, contact tuning, real-time loading, and rendering pipelines. * Have led projects that bridge sim and robotics, such as teleoperation data collection, hardware-in-the-loop regression tests, or digital twin efforts. * Experienced building robust tools and processes — CI for sims, asset validation pipelines, or distributed GPU workloads — and care about documentation and ergonomics that let others be productive. * Can collaborate broadly: you want to work directly with researchers to design experiments, with SWE/RE to ship tooling, and with ve