Research Scientist, SAMBA, DeepMind
We sent a six-digit code to . Enter it below — or use the link in the same email.
Or click the link in the same email — either works.
Enter the email address on your account and we'll send you a link to set a new password.
Remembered it? Sign In
Looking for a research scientist job in New York? Choose from 36 open roles at 19 employers. Explore real salary data and skills breakdown across remote, hybrid and on-site roles.
Free. No credit card needed.
The work is largely focused on advancing large language models and AI agent systems, with a strong emphasis on evaluation and safety. Several roles involve designing evaluation frameworks and benchmarks to surface failure modes in frontier models, as in the Machine Learning Research Scientist, Evaluations posting at Scale AI. Others centre on making AI systems safer and more controllable: the Research Scientist, Safety Post Training role involves post-training pipelines and interpretability techniques, while the Research Scientist, AI Controls and Monitoring role covers real-time behavioural tracking and red-team simulations. Agent robustness features in several titles, and the Multimodal Alignment, Safety, and Fairness role at Google DeepMind extends this to cross-modal systems. A smaller number of postings lean toward pre-training and tokenisation research. One posting, Senior Manager, Research Scientist at Scale AI, also includes team leadership, roadmap definition, and mentoring responsibilities.
The skills data reflects how often terms appear in posting descriptions across this sample of listings, not formal requirements. PhD and evaluation are the most widely mentioned terms, followed closely by LLMs and a publication record. Python and reinforcement learning appear in more than half the postings. PyTorch and statistics appear in roughly a quarter. Less common mentions include distributed training, fine-tuning, GPU experience, and NLP. Presence of a term does not confirm it is mandatory for any individual role.
The employers' own advertised ranges, for individual roles at different levels — not an average and not a market rate.
Most of the postings on this page are listed as onsite, making that the dominant mode in this set. A small number are hybrid, and a minority are fully remote. No postings had unknown work-mode information, so every listing has a stated arrangement. Candidates who need flexibility should read each posting carefully, as the onsite-heavy distribution means hybrid and remote options are limited within this particular snapshot. Specific office locations are not always detailed in the summaries shown.
Jobs checked 1 hour ago. · Guide reviewed 14 September 2026.
Search by role, company, or anything a posting mentions.
