Create and maintain a flexible, modular simulation framework designed for large-scale AI training applications.
Use advanced physics and simulation engines to model realistic physical behaviors, including rigid body motion, intricate shapes, and varied material characteristics.
Build and fine-tune detailed simulation models for multiple sensor types, ensuring realistic output through quantitative validation against real-world data.
Design efficient, scalable systems for generating large volumes of synthetic data offline, integrated directly into machine learning training processes.
Identify potential failures, risks, and inefficiencies in data generation workflows, and establish monitoring systems to reduce discrepancies between simulation and reality.
Develop adaptable asset creation pipelines to support diverse hardware setups, sensor arrangements, and changing operational environments.
Build core simulation infrastructure to enable future advancements in domain randomization, covering spatial, visual, and physical variations as AI models grow in sophistication.

Competitive salary & equity

Design, develop, and maintain a modular, extensible, and customizable simulation platform tailored for scalable AI applications.
Leverage state-of-the-art simulation and physics engines to accurately model realistic physical interactions, including rigid body dynamics, complex geometries, and diverse material properties.
Develop and calibrate high-fidelity simulation models for diverse sensor modalities. Establish rigorous, quantitative validation metrics to ensure simulated perception accurately matches real-world ground truth.
Architect scalable pipelines for high-throughput offline synthetic data generation, ensuring seamless integration with machine learning training loops (e.g., Reinforcement Learning, Computer Vision).
Proactively identify and mitigate risks, failure modes, and bottlenecks for synthetic data generation pipeline. Define and monitor metrics to continually measure and minimize the sim-to-real gap.
Implement highly configurable asset pipelines to support a wide variety of hardware topologies, sensor configurations, and dynamic operational layouts.
Architect foundational systems capable of supporting next-generation, contact-rich interactions. You will lay the groundwork for our future efforts in advanced domain randomization—expanding spatial, visual, and physical parameters—as our AI models scale in complexity over time.

UnitX is hiring a Simulation Engineer