Research Scientist, Vision Benchmarks

Roboflow; San Francisco (CA); On-site; USD 250,000 - 450,000; Full time

Benefits offered by this job

Stipends

Health coverage

Unlimited PTO

Parental leave

Job summary

Roboflow is seeking a Research Scientist for Roboflow Labs to advance benchmarks, models, and real-world datasets. You will run failure analyses that pinpoint where a model breaks on tasks, translate findings into dataset and evaluation specs, and publish benchmarks that guide the research agenda.

Design evaluations across frontier vision-language, multi-modal, robotics, and coding models; perform RL and fine-tuning; curate data; contribute to open models like RF-DETR; collaborate with frontier

Qualifications

• PhD or equivalent research experience in computer vision, multi-modal learning, machine learning, or a closely related field.
• Hands-on experience with post-training: fine-tuning, RL, evaluation design, or data curation for large models.
• Strong engineering fundamentals — you write the code for your own experiments and can hand it to an engineer to productionize.
• Clear written communication; you can explain a failure mode to a lab researcher and a data spec to an annotator.
• Nice to have: frontier lab experience or as data/evaluation partner to one; vision-language or robotics learning knowledge.

Responsibilities

• Design datasets, evaluations, and RL environments for frontier vision-language and multi-modal models.
• Run structured failure analyses against real-world task distributions and write remediation specs.
• Own the research behind new benchmarks and contribute to Roboflow’s open models.
• Validate data quality empirically and ensure datasets come with evidence they work.
• Collaborate with frontier labs as a technical peer and translate model roadmaps into data programs.

Skills

PhD in CV

Post-training methods

Engineering fundamentals

Technical communication

frontier-lab experience

Education

PhD in computer vision or related

Job description

Roboflow is seeking a Research Scientist for Roboflow Labs to advance benchmarks, models, and real-world datasets. You will run failure analyses that pinpoint where a model breaks on tasks, translate findings into dataset and evaluation specs, and publish benchmarks that guide the research agenda.

Design evaluations across frontier vision-language, multi-modal, robotics, and coding models; perform RL and fine-tuning; curate data; contribute to open models like RF-DETR; collaborate with frontier