Fast, Scalable Vision Inference Engineer

US Health Partners, LLC; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Benefits offered by this job

Competitive compensation and equity

401k

Healthcare (Silver PPO Medical, Vision

Lunch and snacks at the office

Job summary

Hedra is a small, highly technical team in San Francisco building state-of-the-art visual models and the infrastructure to deploy them at scale. We seek an Inference Optimization Engineer to push model performance at inference time, working at the boundary between research and systems.

You’ll collaborate with researchers to run new architectures efficiently on modern hardware, exploring novel optimization techniques and scalable execution across accelerators.

Qualifications

• Deep technical ability in efficient ML inference, ML systems, GPU computing.
• Strong programming fundamentals in Python, C++, or another systems language.
• Experience with PyTorch, CUDA, Triton, TensorRT, vLLM, SGLang, or comparable technologies.
• Ability to reason across abstraction layers and collaborate across research and engineering boundaries.

Responsibilities

• Work with research scientists and engineers to optimize inference for new visual models.
• Profile architectures and workloads to identify bottlenecks in compute, memory, and communication.
• Develop and implement approaches to reduce latency and improve throughput.
• Explore optimizations including quantization, sparsity, caching, and alternative execution strategies.
• Build or optimize GPU kernels using CUDA, Triton, or similar tools.
• Optimize execution across single-GPU, multi-GPU, and multi-node setups.
• Study interaction between model design and hardware to unlock performance gains.
• Create benchmarking and profiling infrastructure to evaluate progress.
• Evaluate new runtimes, compilers, and accelerator hardware.

Skills

Python

C++

ML systems

GPU computing

Profiling ML workloads

Tools

PyTorch

CUDA

Triton

TensorRT

vLLM

Job description

Hedra is a small, highly technical team in San Francisco building state-of-the-art visual models and the infrastructure to deploy them at scale. We seek an Inference Optimization Engineer to push model performance at inference time, working at the boundary between research and systems.

You’ll collaborate with researchers to run new architectures efficiently on modern hardware, exploring novel optimization techniques and scalable execution across accelerators.