ML Engineer, Inference Optimization

About the role

Build AI is a data hyperscaler focused on Physical AI, integrating hardware, manufacturing, logistics, data collection, and model training to accelerate physical labor datasets. The ML Engineer – Inference Optimization will lead efforts to make inference cheaper, faster, and scalable, directly impacting the company’s compute costs.

Key responsibilities

• Own inference performance metrics such as latency, throughput, and cost per token, frame, or job.
• Reduce compute usage by optimizing kernels, batching, quantization, compilation, serving, and hardware utilization.
• Profile pipelines with tools like Nsight or PyTorch Profiler, identify bottlenecks, and deliver fixes.
• Collaborate with research and product teams to ensure models remain accurate while being affordable at scale.
• Build serving and evaluation pipelines that expose true inference costs during experiments.
• Treat cost as a first‑class metric throughout development.

Required profile

• Strong ML/systems engineering background with hands‑on inference optimization experience (serving, compilers, CUDA/kernels, quantization, etc.).
• Proficiency in Python and in C++ or Rust for performance‑critical code.
• Ability to think in dollars and tokens/frames per second rather than only accuracy.
• Familiarity with PyTorch or JAX and profiling tools.
• Comfort working in a small research team under cost pressure.

Required skills

• Python
• C++
• Rust
• CUDA
• Kernel development
• Quantization
• PyTorch
• JAX
• Profiling tools (Nsight, PyTorch Profiler)
• TVM
• MLIR
• TensorRT

What we offer

• Competitive salary and comprehensive medical, dental, and vision coverage.
• $500 monthly credit for waiving medical benefits.
• Housing subsidy of $2k per month for nearby residents.
• Relocation support to San Francisco or Shenzhen.
• Wellness benefits covering fitness, mental health, and more.
• Daily lunch and dinner on‑site, unlimited compute budget (subject to ROI), and unlimited Codex and Claude credits.