Get Started
Hunting for you, 24x7Get the iPhone app

Your hunt runs in your pocket, and tells you the moment a strong match opens.

Download on theApp Store
RemoteJobsOnePosted in November 2026

Member of Technical Staff, Coding Agent Research & Evaluations – Remote (Worldwi

Who we'd reach at RemoteJobsOneFound and reached by Power Apply
Best leadHiring manager

Leads the team this role joins

Team director

Heads the group this role sits in

Recruiter

In-house, owns the opening

Power Apply

From your own Gmail and LinkedIn.

About the role

This is a full-time, fully remote research role for people who will design benchmarks and evaluation methods for coding agents and frontier coding models. It suits engineers with 3 or more years of experience who can combine software engineering with model evaluation, and it is open only to candidates in the listed countries. Much of the work is building measurement tools, datasets and evaluation pipelines, then turning the findings into improvements.

What you’ll do

  • Design and own evaluation frameworks for coding agents, including benchmark specifications, scoring methods and rubrics.
  • Lead research on how coding models perform across varied software engineering tasks.
  • Create datasets, reference examples and evaluation protocols for assessing coding systems.
  • Study model behavior and failure patterns, and turn them into training and evaluation changes.
  • Build tooling for large-scale experiments, data generation, review workflows and evaluation pipelines.
  • Set standards for reproducibility and measurement quality in coding-agent assessment.
  • Work with researchers and applied AI teams on experiments, and contribute to technical reports and client-facing studies.

What they’re looking for

  • A strong software engineering background, with Python, C++ or comparable languages.
  • 3 or more years in software engineering, machine learning, AI research, evaluation or related fields.
  • Experience designing, reviewing or validating benchmarks, coding tasks or evaluation methods.
  • Familiarity with large language models, coding agents, reinforcement learning or model evaluation.
  • A proven ability to build tooling, automate workflows and run systematic experiments.
  • Strong analytical skills for investigating model behavior.
  • Clear written and verbal communication for technical and general audiences.
  • Comfort working in fast-moving research settings with ambiguity and shifting priorities.

Nice to have

  • Experience with frontier AI systems, coding agents or model evaluation research.
  • Interest in how data, evaluations and feedback shape model capability.
  • A record of driving ambiguous technical or research projects from start to finish.
  • Experience designing benchmarks or datasets for machine learning at scale.
  • Familiarity with agentic workflows, tool use, reinforcement learning or post-training methods.
  • Publications, open-source contributions or technical leadership in AI, machine learning or software engineering.

What’s on offer

  • Base salary stated as $200,000 to $260,000 USD in the compensation notice.
  • Equity for all employees, with possible performance-based bonuses.
  • Health insurance premium reimbursement of up to 100%.
  • Paid time off and a 401(k) plan with company match.
  • Fully remote working arrangement.

Skills

Questions about this role

Is this role remote?

Yes. It is a fully remote, full-time position.

How much does it pay?

The compensation notice gives a base salary of $200,000 to $260,000 USD. A pay line at the top of the listing shows $600,000 to $1,300,000 per year, so confirm the figure with the employer.

Who can apply?

The listing names eight eligible countries: Canada, France, Germany, India, the Netherlands, the Philippines, the United Kingdom and the United States. Elsewhere it refers to candidates in 55 countries, so check eligibility before applying.

How do I apply?

Apply with your profile, then complete the application through the link emailed to you. The application is only considered once that final step is done.