Member of Technical Staff - Model Optimization and Inference

Company: Confidential venture-backed technology company
Location: Seattle, WA (in office 5 days per week; relocation assistance available)
Compensation: $250,000 - $350,000 + equity
Employment Type: Full-time
Visa Sponsorship: Visa sponsorship available

The client builds photorealistic, real-time AI avatars with emotional intelligence: a full-duplex audiovisual system that can listen, speak, react and respond like a real person. The research team includes PhDs from MIT, UW, Oxford, CMU and Johns Hopkins.

Founded in 2024, the client has raised $60M and has about 25 people.

The client is hiring an experienced ML Infrastructure/Systems Engineer (2+ years) to own end-to-end inference optimization across LLMs, audio models and diffusion components, with a focus on latency, throughput and cost.

• Own end-to-end inference optimization across the model stack.
• Implement and tune KV cache strategies for long-context conversations.
• Evaluate, deploy and extend serving frameworks such as vLLM, SGLang and TensorRT-LLM.
• Profile and benchmark latency and throughput and remove bottlenecks.
• Accelerate diffusion inference and apply quantization techniques (INT8, INT4, GPTQ, AWQ).
• Build internal tooling that makes optimization work faster and more rigorous.

• 2+ years building and maintaining production ML systems
• Designing scalable infrastructure from scratch
• Track record optimizing latency, throughput and cost
• Debugging distributed systems

• Video or audio model experience
• CUDA kernels and low-level optimization
• Real-time video streaming (WebRTC)

HSA with about $2,000 annual company contribution, 15 days PTO plus public holidays and a company-wide office closure week.

Kubernetes, Terraform, Python, Rust, Go, Dagster, Ray, Airflow, WebRTC, vLLM, Triton Inference Server, TensorRT

Before you apply: get a free recruiter-readiness score for your resume at CareerStanding about a minute, no signup, and your resume stays in your browser.