Founding Reinforcement Learning Researcher

Overview

VectraSim is building an AI‑native, end‑to‑end engineering simulation platform that aspires to replace weeks of expert labor with near‑real‑time turnaround. Behind the scenes, a hierarchy of specialist AI agents orchestrates the traditionally manual steps and replace weeks of expert labor with near‑real‑time turnaround. Ultimately, VectraSim evolves into the default "autonomous engineer" for physical product development, powered by our own large‑scale physics foundation models and custom GPU solvers. We are backed by the best including Fifty Years VC, Jeff Dean (head of Google AI), and Wojciech Zaremba (OpenAI cofounder).

The Role

We are looking for a Founding Reinforcement Learning Engineer to help lead core R&D and work directly with iconic engineering companies to prototype, refine, and launch the initial versions of our product. This role provides a unique opportunity to shape our early technical vision, guide the design of scalable methods, and make immediate impact on real, high-stakes engineering challenges.

Responsibilities

• Develop RL methods for engineering design optimization
• Build and adapt foundation models from simulation data
• Integrate PDE-based solvers with large-scale ML systems
• Design agent hierarchies for autonomous engineering workflows

Required Skills

• Expertise in foundation models for scientific and engineering applications
• Research in dynamic and few-shot RL (e.g., meta-learning, adaptive policies)
• Experience with LLM-integrated RL for planning and policy adaptation
• Familiarity with PDEs and numerical methods