Member of Technical Staff, Forward Deployed

present of AI; San Francisco (CA); On-site; USD 140,000 - 190,000; Full time

Job summary

Wafer is seeking a Member of Technical Staff, Forward Deployed in San Francisco. This customer-facing engineering role focuses on running benchmarks on a customer's workload, explaining the results to their engineers, and getting them live in production.

The candidate will need to scope half-defined work and deliver concrete answers rather than questions. The role offers hands-on exposure to LLM inference and serving stacks, such as vLLM, SGLang or TensorRT-LLM, and running open models on AMD

Qualifications

• Experience shipping and running production systems.
• Ability to read traces and profiles to diagnose problems.
• Familiarity with LLM inference serving stacks (vLLM, SGLang, TensorRT-LLM).
• Proven ability to own technical relationships with customers.

Responsibilities

• Run benchmarks on a customer’s workload and explain results to their engineers.
• Help customers get deployments live in production and maintain health.
• Scope ambiguous work and deliver clear answers rather than asking excessive questions.
• Provide feedback to engineers building the AI engine, kernels, and agents.

Skills

Production systems

Trace and profiler debugging

LLM inference serving stacks

Technical customer ownership

Tools

vLLM

SGLang

TensorRT-LLM

Job description

Member of Technical Staff, Forward Deployed

Customer & Solutions · San Francisco · FullTime · Mid level

This role is a customer facing engineering job: running benchmarks on a customer's own workload, explaining the results to their engineers, and getting them live in production. A strong candidate has shipped and run production systems, can read traces and profiles to find problems alone, and has handled a technical customer relationship. The work arrives half defined, so the person needs to scope it and return with an answer rather than a list of questions.

How it fits the AI era

The tuning itself is largely done by performance engineers and automated optimization agents, so the human job is making that tuning land for each customer and keeping it healthy afterward. AI changes the pace: benchmarks and monitoring can be partly automated, but judging whether a result is credible and defending the method to a skeptical team stays with the engineer. The role also feeds what customers need back to the people building the engine, kernels and agents.

How it builds AI experience

The person would gain hands on exposure to LLM inference and serving stacks such as vLLM, SGLang or TensorRT-LLM, and to running open models on AMD and Nvidia chips. Working beside optimization agents on real customer traffic builds skill in checking automated tuning against production latency and error targets. Prior serving experience is a plus, and the rest can be built by operating these stacks on live workloads.

Worth showing

• Production systems shipped and run
• Trace and profiler debugging
• LLM inference serving stacks
• Technical customer ownership

This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.

About Wafer

An AI inference provider that uses AI agents to tune GPU code, which it says lets open models run faster on AMD and Nvidia chips.