Reinforcement Learning for Vision Language Models

Overview

Microsoft AI is looking for a Member of Technical Staff to advance the capabilities of frontier vision-language models through reinforcement learning. Within our Large Scale Reinforcement Learning team, you will help models better understand visual information, reason from visual evidence, and provide accurate, useful responses to complex multimodal questions.

Our work spans visual perception and grounding, document and chart understanding, image-to-code generation, and spatial and physical reasoning from visual inputs. We aim to develop capabilities that generalize to unfamiliar problems and real-world use, rather than simply improving benchmark scores.

You will own projects across the research loop: identifying capability gaps, developing training tasks and data, designing reliable reward signals, running large-scale experiments, and turning the results into better models. This is a hands-on research and engineering role, combining scientific judgment with strong implementation skills. You will work closely with researchers and engineers across reinforcement learning, multimodal modeling, data, infrastructure, evaluation, and safety.

Responsibilities

• Advance visual understanding and multimodal reasoning through reinforcement learning, translating capability gaps into concrete research hypotheses and measurable improvements.

• Develop targeted training tasks and data, including synthetic data, to address these gaps.

• Design and validate rewards, verifiers, rubrics, and model-based judges for both verifiable and open-ended visual tasks, detecting and mitigating reward hacking.

• Build evaluations and conduct controlled experiments, ablations, and detailed failure analysis.

• Collaborate with scaling, multimodal, data, and safety teams to integrate improvements while preserving broader model capabilities.

Required Qualifications

• Bachelor's degree in Computer Science, Machine Learning, Mathematics, or a related technical field and 4+ years of relevant research or engineering experience;

  • OR master's degree in a related field and 2+ years of relevant research or engineering experience;

  • OR PhD in a related field.

• Experience developing, training, or evaluating deep learning models, with demonstrated depth in at least one of model post-training, reinforcement learning, or vision-language modeling.

• Ability to formulate hypotheses, design informative experiments, analyze results, and debug model or training behavior.

• Ability to take end-to-end ownership of technical projects and collaborate effectively across research and engineering disciplines.

• Proficient programming skills in Python and PyTorch (or similar).

Preferred Qualifications

• Bachelor's degree in Computer Science, Machine Learning, Mathematics, or a related technical field and 8+ years of relevant research or engineering experience;

  • OR master's and 6+ years of relevant research or engineering experience;

  • OR PhD and 4+ years of relevant research or engineering experience.

• Experience training models via reinforcement learning across distributed environments.

• Demonstrated ability to build models for complex visual reasoning tasks such as visual grounding, document and chart understanding, image-to-code generation, and spatial/physical reasoning.

• Experience developing synthetic data, reward models, automated verifiers, or model-based evaluation systems, including assessing their reliability and failure modes.

• A record of meaningful research or engineering contributions, demonstrated through publications, open-source work, deployed systems, or other substantial technical projects.

Software Engineering IC5 - The typical base pay range for this role across United Kingdom is £ 91,800 - £ 137,600 per year.