Job Summary:
Open Orion is an early stage research and design company aiming to enhance engineering through AI and simulations. They are seeking a strong ML Engineer to build agentic AI systems that reason about geometry and physics, develop generative models, and create systems for complex structured data. This role offers direct influence over the company's technical direction and product development.
Responsibilities:
• Design and build agentic AI pipelines systems that can autonomously plan, execute, and adapt across multi-step workflows
• Develop generative models for structured, constraint-satisfying outputs (not just text and images think geometry, configurations, designs)
• Build surrogate models that learn to approximate expensive computational processes from data
• Create reasoning systems that can classify, manipulate, and make decisions over complex structured representations
• Integrate ML models into production pipelines for optimization and design space exploration
• Develop and maintain open-source tools that push the state of the art
Qualifications:
Required:
• Strong ML/AI fundamentals deep learning, transformers, generative models, reinforcement learning
• Experience building agentic systems or multi-step AI pipelines that plan, execute, and adapt
• Proficiency in Python and at least one ML framework (PyTorch, JAX, TensorFlow)
• Experience with structured prediction, 3D data, or scientific/engineering ML (not just NLP/vision)
• Comfort working in ambiguous problem spaces where the right approach isn't obvious
• Strong software engineering habits you write code that other people can read, test, and build on
Preferred:
• Experience with 3D deep learning point clouds, implicit representations, graph neural networks
• Familiarity with computational geometry concepts meshes, BREP/NURBS, surface discretization
• Exposure to CAD/CAE workflows, simulation tools, or geometry kernels
• Background in surrogate modeling or physics-informed neural networks
• Experience with differentiable simulation or differentiable rendering
• Hands-on experience with CFD, FEM, or mesh generation
• Contributions to open-source scientific computing projects
• Aerospace, mechanical, or civil engineering domain knowledge
Company:
Founded in , the company is headquartered in , , with a team of 2-10 employees. The company is currently Early Stage.