Save application

Toptal

3 days ago

ML Inference Engineer – Terminal-Bench Task Designer (vLLM / SGLang / TensorRT-LLM)

Remote · Anywhere in the World Anywhere in the World

pythondockerpytorch artificial intelligencedockergitgraphics processing unitllm

remoteok

Match & cover letter

Create a profile to see your match and get a cover letter for this role.

Create profile and match

Source: remoteok

Never pay or share passwords or codes with an employer. How to spot a fake job

AI TL;DR

You will design expert-level tasks for AI agents based on real ML inference work. This role allows for remote work and is compensated per task completed.

AI match

Clear role · 80/100 against the market band on this page.

Interview prep for this job

  1. What would you ship in the first month as ML Inference Engineer – Terminal-Bench Task Designer (vLLM / SGLang / TensorRT-LLM) at Toptal?
  2. Which tradeoff in the description would you ask about first?
  3. How do you explain a production incident to this team?
  4. What would you measure in the first two weeks?
  5. Which part of your last project maps to this stack?

Five questions are free. The rest are Pro. See Pro

Similar roles

Job description

You will design expert-level tasks for AI agents based on real ML inference work. This role allows for remote work and is compensated per task completed.

Details

  • Remote work from anywhere in the world
  • 3+ years of hands-on ML inference or serving experience
  • Availability to start by early November 2026

The work

  • Design expert-level Terminal-Bench tasks from real ML inference work
  • Define environments and systems for each task to ensure realism
  • Write and verify working reference solutions for tasks
  • Build automated checks to validate task results