LA Startup VmaxAI Hires DeepMind Talent
Emerging LA-area startup VmaxAI has hired former Google DeepMind engineer Geoffrey Bradway, an expert in reinforcement learning. The move signals that local startups are successfully attracting top-tier AI talent from big tech for specialized roles like automated environment design.
VmaxAI, officially a San Francisco-based startup, was founded in 2025 by Matthew Sargent and Augustine Mavor-Parker, both of whom hold PhDs in reinforcement learning. The company is in its seed stage and has secured initial funding from South Park Commons, a venture capital firm known for backing early-stage tech companies. Publicly, the company has a reported team size of 2 to 10 employees. The firm's technical focus is on automating the generation of reinforcement learning (RL) environments. This involves creating complex, scalable simulations where AI agents can learn and be tested, a critical bottleneck in the development of more advanced AI systems. VmaxAI aims to use large language models to convert proprietary data into these specialized RL environments for long-horizon tasks. This move into automated environment design represents a significant step beyond traditional prompt engineering, aiming to build entire "kitchens" for AI rather than just giving them "recipes." The goal is to create dynamic, interactive ecosystems where an AI can learn and operate autonomously within defined safety and ethical boundaries. The hiring of talent with expertise in reinforcement learning from major AI labs is a key part of VmaxAI's strategy as they build out their research and development capabilities in this specialized field. The company is actively recruiting for roles focused on the automated design of RL environments and applied reinforcement learning.
