Senior Domain Expert: Multi-Discipline AI Research Program

What you'll do

This is research-and-evaluation work, not production engineering. The ad says you will define what "correct" and "excellent" look like on problems you already know deeply, to help evaluate how frontier AI models reason about hard, expert-level problems. Concretely: design realistic technical scenarios and problem sets within your track, author expert-level reference solutions and grading rubrics, and evaluate AI-generated output for correctness, depth and domain judgment.

The seven tracks

You must fit one of these profiles to be considered.

• Senior Software/Systems Engineer: 8 to 12 years; bachelor's (master's preferred)
• Research Scientist, Formal Methods / Computational Science: 10+ years; PhD required
• ML Research Engineer, Inference & GPU Kernels: 6 to 10 years; PhD, or master's with a strong research record
• Enterprise Operations / Domain Analyst (supply chain, finance, compliance): 8 to 12 years; bachelor's (professional certifications a plus)
• Security Engineer, Cryptanalysis / Reverse Engineering: 6 to 10 years; bachelor's (master's preferred)
• Mechanical / Hardware Design Engineer (CAD, RTL, robotics): 8 to 12 years; bachelor's (PE license or master's a plus)
• Creative Technologist, Audio / Design / Linguistics: 6 to 10 years; bachelor's (master's a plus)

Beyond the track bar: currently or recently active and hands-on in the field, strong written communication (you will author technical explanations and structured feedback), and comfort with independent, asynchronous, remote work. Preferred: publications, open-source or recognised work product, experience reviewing others' work, and track bonus signals such as SGLang, vLLM, Mamba, TensorRT-LLM or GPU kernels for ML, Lean or Coq for science, CTF experience for security, RTL/HDL for hardware, and audio engineering or linguistics for media.

What it pays

The board shows $135/hour. The ad says compensation is "scaled to track and seniority (see below)" but does not give a per-track figure or range, so this listing records $135/hour as a single figure and the real rate for your track may differ. Payment terms are not published on the listing; see AfterQuery for how payouts generally work.

Worth knowing

Good:

• High-seniority work that uses expertise rather than routine labelling
• Seven tracks, so it is open beyond pure software
• Remote, flexible and asynchronous around existing work

Less good:

• The ad warns that candidates who meet fewer than 70% of the stated qualifications may be flagged for misrepresenting their professional experience, which could affect eligibility for future project staffing, and says it conducts background checks
• The experience bars are high, and the science track requires a PhD
• Pay by track is not published, despite the "see below" note
• The ad names no weekly hours and no hiring steps; AfterQuery's help centre describes an unpaid, one-attempt assessment and ID verification

About this listing

Read from the AfterQuery Experts board on 18 October 2026, in the Miscellaneous department. Tracks, bars and the review warning are AfterQuery's own text. See AfterQuery.