About the team:
The Software Engineering Team builds and operates the systems that power Formic's Robotics-as-a-Service platform.
Engineering focuses on ensuring deployed systems are observable, resilient, and remotely diagnosable at scale. The team builds production-grade edge and cloud systems that support reliable data collection, remote troubleshooting, live video streaming, and continuous system improvement.
This work directly impacts fleet uptime, service efficiency, and customer outcomes by ensuring Formic's monitoring and control infrastructure remains scalable, reliable, and continuously evolving.
In this role you will:
- Design and train learning-based manipulation systems for humanoid and mobile manipulation platforms
- Develop and maintain high-fidelity digital twins using Isaac Sim, MuJoCo, or similar tools
- Implement and evaluate approaches such as:
- Action Chunking with Transformers (ACT)
- Diffusion Policies
- Behavior Cloning at Scale
- Vision-Language-Action (VLA) models
- Latent action or hierarchical skill models
- Work with modern foundation models for robotics (e.g., Pi0, Gemini ER 1.6 or similar), including adapting, fine-tuning, and deploying them for real-world tasks
- Contribute to development of a Universal Manipulation Interface (UMI) abstraction layer
- Build teleoperation-to-training data pipelines
- Design sim-to-real transfer strategies including domain randomization and system identification
- Evaluate policy robustness, generalization, and real-world performance
- Work closely with perception and controls teams to ensure stable closed-loop visuomotor policies
- Deploy, tune, and iterate on models running on real robotic systems in production environments
What makes you a great fit:
- Experience training embodied AI policies for real robots
- Familiarity with transformer-based action models (e.g., ACT)
- Experience with diffusion policies or other generative control methods
- Experience working with or adapting large-scale models (e.g., Pi0, Gemini ER 1.6, or similar VLA / multimodal models)
- Ability to deploy, fine-tune, and optimize models for real-world robotic systems (latency, robustness, reliability)
- Strong understanding of sim-to-real challenges
- Experience working with multi-modal inputs (vision, proprioception, language)
- Proficiency in Python and deep learning frameworks (PyTorch, JAX, etc.)
- Experience integrating learned policies with real-time control systems
- Strong experimental design, evaluation, and debugging skills