Member of Technical Staff - Low Level & Kernels Capabilities
Core
Building reinforcement learning (RL) environments at the lowest layers of the stack (GPU/accelerator kernels, vector ISAs, FPGA) to address brittleness in frontier models for real-world tasks.
Role type
Senior IC machine learning engineer (low-level systems & kernels)
Builds
RL environments, scoring infrastructure, and sandboxing mechanisms for model evaluation
Domain
AI/ML research, hardware systems, reinforcement learning
Deliverable
production ML models
Required skills
C/C++, CUDA, assembly, Python (production quality), hardware-aware coding, kernel optimization, adversarial mindset, LLM interaction
Preferred skills
FPGA/HLS/RISC-V/SIMD/AVX/TPU expertise, HPC/heterogeneous clusters, hardware design (RTL/HDL), compilers (MLIR/LLVM/Mojo/Triton), formal verification, competitive programming, RL environment construction
Technologies
CUDA, gem5, cuBLAS, FFTW, OpenSSL, MLIR, LLVM, Mojo, Triton, Lean, Coq, SMT
Responsibilities
Design and build low-level RL environments targeting specific models and difficulty distributions; develop deterministic, ungameable scoring and sandboxing mechanisms; optimize kernels against profilers; select domains based on hardware features and benchmark gaps
Seniority
Senior, hands-on IC
