Microsoft AI
Multimodal Safety
Get notified of more jobs like this · No spam, ever
How your CV stacks up
Upload your CV to see how well it fits this job role
?%
About Microsoft AI
Microsoft AI is building AI systems and products that empower people’s lives. Our work is driven by a community of brilliant, interdisciplinary minds working across frontier model development, product engineering, and responsible AI. Within Microsoft AI, the Safety team develops the training methods, evaluations, runtime safeguards, monitoring, and infrastructure needed to make advanced AI systems safer, more reliable, and more useful. Our work spans text, multimodal, and agentic systems and is developed in close partnership with other Research teams, Production Inference, Security, Responsible AI, Microsoft product organizations, and external partners and customers.
About the Role
We are looking for a Member of Technical Staff to advance safety for generative-media models and products, with a focus on image generation and image editing. You will develop model-based classifiers, runtime guardrails, evaluations, and data that help these systems produce creative and useful content while respecting safety policy. You will work across research and production boundaries: identifying emerging failure modes from evaluations and real-world use, reproducing and diagnosing them, improving guardrails and policies, and partnering with product and platform teams to deploy and monitor safeguards at scale.
This role is a strong fit for an ML or software engineer who combines an interest in generative image systems with rigorous experimental judgment and a practical understanding of safety policy, creative quality, overblocking, latency, and real-world deployment.
Responsibilities
- Develop and improve learned classifiers, policy models, prompt-based safeguards, and other runtime guardrails for MAI's multimodal models.
- Build high-quality training and evaluation datasets from synthetic generation, red teaming, product feedback, and real-world traffic while maintaining appropriate controls.
- Design evaluations for harmful or deceptive generated media, unsafe image transformations, policy evasion, adversarial inputs, and emerging image and audio product use cases across both request and output paths.
- Turn incidents and partner feedback into repeatable test cases and regression suites.
- Diagnose failures, then translate findings into durable model, policy, or product improvements.
- Monitor deployed safeguards across traffic sources, customers, models, and product configurations.
- Partner with Production Safety and serving teams on Azure-based deployments, capacity planning, and guardrails integration.
- Partner with Safety Alignment, Safety Engine, model, product, policy, privacy, and Responsible AI teams to adjudicate ambiguous cases and define end-to-end safety strategies.
- Contribute reusable guardrail libraries, documentation, deployment and monitoring runbooks, technical standards, and operational practices for generated-media safety.
Reasons to use Rodeo
Start with a chat, not a search bar
Grad scheme, placement, apprenticeship? Not sure what you want yet — that's fine. Your agent talks it through with you and turns "I have no idea" into a shortlist.
It searches the market for you
Every day your agent scans the market matching roles against what actually matters to you, not just keywords on a CV.
Only hits
No noise. No "maybe this fits." Just roles with a clear explanation of why they're right — and where to focus when applying.
Required Qualifications
- Bachelor’s degree in Computer Science, Engineering, a related technical field, or equivalent practical experience.
- Experience developing, adapting, or evaluating modern machine learning systems.
- Strong programming skills in Python.
- Ability to design experiments, define metrics, run ablations, and make evidence-based technical decisions.
- Experience building data pipelines, including dataset construction and annotation.
- Experience developing reproducible ML workflows using version control, configuration management, automated testing, and experiment tracking.
- Ability to collaborate across research, engineering, product, and policy disciplines and communicate technical tradeoffs clearly.
Preferred Qualifications
- Experience working with diffusion models, autoregressive media models, audio-generation systems, or other architectures used for image or audio generation and editing.
- Experience building image or audio safety classifiers, content-understanding systems, trust and safety models, policy engines, or input and output guardrails for user-facing products.
- Familiarity with generated-media safety risks such as adversarial inputs, unsafe generation or editing, harmful-content transformation, impersonation or deceptive media, privacy violations, and policy evasion.
- Experience with Azure or another major cloud platform, including Azure AI services or Azure Machine Learning, model APIs, GPU capacity planning, versioning, monitoring, and rollback.
- Experience with red teaming, adversarial evaluation, interpretability, robustness, privacy, security, policy development, or responsible AI.
Get help with your application
Your very own career expert that helps elevate your application to the next level.
Compensation
Software Engineering IC4 - The typical base pay range for this role across the U.S. is USD $119,800 - $234,700 per year. There is a different range applicable to specific work locations, within the San Francisco Bay area and New York City metropolitan area, and the base pay range for this role in those locations is USD $160,200 - $261,000 per year.
Certain roles may be eligible for benefits and other compensation. Find additional benefits and pay information here: https://careers.microsoft.com/us/en/us-corporate-pay
This position will be open for a minimum of 5 days, with applications accepted on an ongoing basis until the position is filled.
Microsoft is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to age, ancestry, citizenship, color, family or medical care leave, gender identity or expression, genetic information, immigration status, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran or military status, race, ethnicity, religion, sex (including pregnancy), sexual orientation, or any other characteristic protected by applicable local laws, regulations and ordinances. If you need assistance with religious accommodations and/or a reasonable accommodation due to a disability during the application process, read more about requesting accommodations.
“It took my CV and asked me questions relevant to understanding what kind of jobs to suggest for me. Suggestions were almost perfect. Jobs were exactly what I’ve been looking for.”
Jessica, London
Location