We're an AI and physical sciences company building state-of-the-art models to accelerate breakthroughs across materials, energy, and beyond. Backed by world-class investors and growing rapidly, we operate at the pace the frontier requires. Our team brings deep expertise, genuine ownership, and an insatiable drive to push the boundaries of what's scientifically possible.
About the RoleYou’ll work alongside some of the world’s leading ML systems engineers, including leaders behind Megatron-LM, SGLang, Liger Kernel, TorchRec, CleanRL, TorchRL, and JAX-MD.
We’re looking for exceptional ML Systems Engineers to build the agentic infrastructure powering our large-scale training, inference, and reinforcement learning. You’ll own critical pieces of the ML systems stack to maximize performance, scalability, reliability, and productivity for both engineers and AI agents.
What You'll DoBuild and optimize large-scale training and reinforcement learning infrastructure while ensuring its correctness
Develop high-performance inference and serving systems
Design distributed runtimes and scheduling systems for complex ML workloads
Build secure and large-scale sandboxing and execution environments
Optimize memory, GPU kernels and communication for maximum throughput and end-to-end efficiency
Improve scalability, reliability, and efficiency across the ML systems stack
Strong systems programming and performance engineering skills
Experience building high-performance ML infrastructure at scale
Ability to own complex technical problems end-to-end
Strong coding ability and engineering judgment, including the ability to work effectively with AI agents to design, implement, test, and debug complex systems
High ownership, fast execution, and a passion for pushing the frontier of AI systems and accelerating scientific discovery
You should have deep expertise in at least one of the following:
Training: Strong experience building, debugging and optimizing large-scale training systems with Megatron-LM. Familiarity with TorchTitan, FSDP, veRL, Slime, or other distributed training systems is a plus.
Distributed Runtime: Strong experience with Ray. Familiarity with Monarch or other distributed execution frameworks is a plus.
Inference: Strong experience with SGLang. Familiarity with vLLM, TensorRT-LLM, or production LLM serving systems is a plus.
Sandboxing: Strong experience with secure execution environments, containers, virtualization, or code sandboxing.
GPU Kernels: Strong experience with CUDA, Triton, CUTLASS, CuTe, or custom GPU kernel development.
GPU Communication: Strong experience with NCCL, NVLink, InfiniBand, RDMA, GPUDirect RDMA, or large-scale communication optimization.
Minimum education: Bachelor’s degree or similar experience
Location: Menlo Park, CA (Soon: San Francisco, too)
Compensation: $250,000-$350,000 base + equity
Visa sponsorship: Yes, we sponsor visas and will do everything we can to assist in this process with our legal support.
We’re building a team of the world’s best — the scientists, engineers, and problem-solvers who don’t just follow the frontier, they define it. If you’re driven to bring AI to life in the physical world and make discoveries that have never been made before, you belong here.
Skills Required
- Experience with large-scale inference infrastructure and production-level serving architecture
- Expertise in low-level systems programming and optimization
- Proficiency in GPU cluster scheduling and orchestration
- Ability to write and optimize CUDA kernels
- Experience in profiling and benchmarking distributed ML systems
- Familiarity with checkpoint management and cloud storage integration
- Experience contributing to open source ML infrastructure projects
- Experience in ML algorithm-infrastructure co-design
What We Do
We're building AI scientists and the autonomous laboratories for them to operate.
