Applied AI, Research Engineer
Looking for a research engineer job in New York? Choose from 32 open roles at 13 employers. Explore real salary data and skills breakdown across remote, hybrid and on-site roles.
Free. No credit card needed.
The roles cluster around building, training, and evaluating large language models, with a strong thread of reinforcement learning running through most of them. Several postings are specifically scoped to a domain—chip design, cybersecurity, software engineering, computer use, healthcare and finance—where the engineer designs RL environments and reward signals tailored to that area, then measures impact on model behaviour.
The Model Evaluations role at Anthropic focuses on turning capability questions into reproducible metrics and maintaining the distributed infrastructure that runs them. The Code RL role blends research and product work: designing training environments, running experiments, and optimising pipelines so Claude can handle real codebases end to end. The Scale AI general agents posting sits closer to production, requiring deployment of agent architectures that combine LLM reasoning, tool use, and memory in enterprise settings. Across the sample, candidates are expected to span research thinking and solid software engineering rather than treating them as separate disciplines.
Skill counts reflect how often a term appears in posting descriptions across all 33 listings, not a stated requirement in every case. Evaluation and LLMs appear in the large majority of descriptions; reinforcement learning and fine-tuning appear in more than half. Python and agent-related skills appear in roughly half. PyTorch, GPU work, and distributed training appear in a meaningful minority. PhD is mentioned in roughly a quarter of postings, and a master's degree in a smaller subset; neither figure tells you whether those are hard requirements or preferred backgrounds.
The employers' own advertised ranges, for individual roles at different levels — not an average and not a market rate.
The clear majority of these postings are listed as onsite, meaning a New York office presence is expected. A small number are listed as hybrid—the Computer Use role at Anthropic, for instance, states at least 25% office time. A further small number are listed as fully remote. No postings have an unknown work mode, so the stated arrangements are explicit for every role on this page. If you are weighing flexibility, check each listing individually, as onsite expectations vary in how strictly they are described.
Jobs checked 4 hours ago. · Guide reviewed 9 September 2026.
Search by role, company, or anything a posting mentions.
