Founding Research Engineer - Reinforcement Learning
San Francisco, CA | In-Person
I’m working with an early-stage, well-funded AI startup in San Francisco that is building the infrastructure and reinforcement learning environments behind the next generation of AI systems.
They’re looking for a Founding Research Engineer (RL) to join a small, highly technical team and work directly on the systems, data, and training infrastructure that enable models to learn from complex, real-world environments.
This is a high-ownership role with the opportunity to have a major impact on the technical direction of the company from an early stage.
What you’ll work on:
• Building and scaling reinforcement learning environments and training pipelines
• Developing infrastructure for model training, evaluation, and experimentation
• Designing large-scale data and ETL pipelines for ML workloads
• Improving the reliability and performance of distributed training systems
• Working across research and engineering to take ideas from experimentation → production
• Helping shape the technical foundation of a rapidly growing AI company
What they’re looking for:
• Strong Python and PyTorch experience
• Experience with reinforcement learning, model training, ML infrastructure, data infrastructure, or evaluation systems
• Strong software engineering fundamentals and the ability to build production-quality systems
• Experience working with large-scale ML workloads or distributed systems
• Comfortable operating in a fast-moving, highly technical startup environment
Bonus points for experience with:
• FSDP, DeepSpeed, or distributed training
• Multimodal models
• Large-scale data/ETL infrastructure
• CUDA, Triton, TensorRT, quantization, or inference optimization
• Building RL environments or evaluation frameworks
This is a great fit for someone who wants to work at the intersection of reinforcement learning,
ML systems, and real-world AI, with significant ownership from day one.
📍 San Francisco - in person
If this sounds like you, apply directly or message me for more information.