ML Research Engineer: PhD students

About Anthology
Anthology Fellows is a community of exceptional engineers and scientists focused on training the next generation of AI models. Fellows use domain expertise to identify where current models fail, and then create high-quality training environments and data to teach AI how to improve on realistic engineering and research workflows.
We work with leading labs to accelerate progress on the hardest problems in software engineering, biology, medicine, materials science, energy, and beyond. Anthology is part of Polymath, which is backed by Base10, Founders Future, Y Combinator, and other incredible investors.
About the role
Anthology is looking for current PhD students with engineering or research experience to create and peer-review technical tasks for training and evaluating AI coding agents.
As an Anthology Fellow, you’ll turn problems from your specialty into tasks with clear objectives, working reference solutions, and verifiers. You’ll also review tasks created by other Fellows, checking their technical correctness, difficulty, and quality.
You must have expertise in one or more of the following:
Kernels: CUDA or Triton, GPU performance, operator fusion, memory optimization, and numerical correctness
Inference: model serving, speculative decoding, quantization, caching, batching, and streaming
Training: pretraining, fine-tuning, distributed training, optimization, and training-pipeline debugging
Data engineering: data pipelines, corpus curation, entity resolution, deduplication, and data quality
Algorithms: graph algorithms, combinatorial optimization, and numerical methods
In this role you will:
Propose original tasks grounded in realistic engineering and research problems.
Write specifications that define the requested outcome, constraints, success criteria, and intended failure modes.
Build reproducible environments, starting code, datasets, and reference solutions.
Develop automated evaluators that check correctness and, where relevant, performance, resource use, or model quality.
Test tasks against AI agent trials and investigate unexpected results.
Peer-review other Fellows’ proposals, implementations, and evaluators within your expertise.
Run reference solutions, check edge cases, and identify ambiguous requirements, unreliable measurements, or shortcuts that could pass evaluation without solving the intended problem.
Give specific, actionable feedback and revise your own tasks in response to review.
You should bring:
Substantial engineering, research, or open-source experience in at least one of the listed domains.
Experience with independently debugging complex systems or implementing technically demanding algorithms.
Strong testing skills, including designing cases that expose subtle correctness failures.
Clear technical writing and the ability to explain why a solution or evaluation is sound.
Comfort with reviewing code, questioning assumptions, and discussing technical trade-offs.
Experience building benchmarks, evaluation frameworks, or reinforcement learning environments is useful. Relevant domain expertise and sound engineering judgment are central to the role.
Compensation
U.S. / Canada / UK / Western Europe: $50–$150/hr
India / Eastern Europe / Latin America: $25–$80/hr
Compensation within each range is based on experience, demonstrated technical depth, domain specialization, location, and the complexity and quality of work the Fellow is able to independently own.
Anthology pays per task. Each task has a pre-approved hour cap, and you invoice for the hours you actually work, up to that cap. Your first task is paid like any other and also helps us gauge fit for the project, but additional tasks aren't guaranteed and depend (in part) on delivered work meeting the quality standards described in the project materials.

Work arrangement
Yes