Senior ML Research Engineer: Inference & GPU Kernels Remote

AfterQuery; Laurel (MD); Remote; USD 138,000 - 234,000; Full time

Benefits offered by this job

Fully remote

Flexible schedule

Competitive hourly pay

Job summary

AfterQuery is building a research lab focused on ML inference and GPU kernel engineering, covering inference serving systems, kernel optimization, and deployment infrastructure for large models. This is research-and-evaluation work, not production engineering, defining what is “correct” and “excellent” on problems you know deeply.

The role expects 6–10 years of experience or equivalent. It’s fully remote, asynchronous, and offers competitive hourly compensation with flexibility in schedule and

Qualifications

• Currently or recently active in the field — hands-on, not purely academic-adjacent
• Strong written communication — you'll be authoring technical explanations and structured feedback, not just doing the work itself
• Comfortable with independent, asynchronous, remote work
• Meets the experience and education bar described in Full Description

Responsibilities

• Design realistic technical scenarios and problem sets within your track
• Author expert-level reference solutions and grading rubrics
• Evaluate AI-generated outputs for correctness, depth, and domain judgment

Skills

Hands-on ML inference

Written communication

Independent remote work

Remote work experience

Education

PhD

Master's with strong research record

Tools

SGLang

vLLM

Mamba/Mamba2

TensorRT-LLM

GPU kernel work

Job description

AfterQuery is building a research lab focused on ML inference and GPU kernel engineering, covering inference serving systems, kernel optimization, and deployment infrastructure for large models. This is research-and-evaluation work, not production engineering, defining what is “correct” and “excellent” on problems you know deeply.

The role expects 6–10 years of experience or equivalent. It’s fully remote, asynchronous, and offers competitive hourly compensation with flexibility in schedule and