We are looking for a senior engineer to solve difficult security problems across Microsoft AI.
Microsoft AI develops and trains first-party models that Microsoft publishes and uses across products including Foundry, Copilot, Microsoft 365, and MDASH. The work spans model development and serving, large-scale compute, shared infrastructure, engineering systems, identity, data, and productivity services - the whole operational system that makes large-scale training, experimentation, and deployment possible. In the last year, MAI evolved from a research lab to a full production model factory. This role is designed for an engineer who has independently built and operated production engineering or security capabilities and is prepared to do so again while the frameworks, operating mechanisms, and secure defaults for frontier AI infrastructure are still being built.
You may come from product or software engineering, security operations, or infrastructure engineering, including detection and response work within those areas. What matters most is builder depth and ability to apply practical risk judgment, not which path you took. You should be able to work through an unfamiliar problem using available evidence, take ownership of a meaningful piece of work, and carry it to a completed, validated result.
This is a broad security-engineering role by design. Initial work may involve building an automated control into our agent framework, improving identity or secrets handling in training systems, defining end-to-end instrumentation and detection for a service, investigating an attack path, hardening a developer workflow, prototyping an AI-assisted security capability, or helping a platform team resolve a systemic risk. You are not expected to begin as an expert in every domain. You will write code, analyze systems and data, build tools and integrations, test controls, and work with engineering teams through implementation and validation. You will use AI routinely for research, coding, analysis, testing, investigation, and automation while remaining responsible for the correctness, security, limitations, and maintainability of the output.
We earn trust with the MAI research community by staying close to the systems we secure, shipping useful improvements rather than only raising findings, making our reasoning and tradeoffs explicit, and telling partners the true state of a risk even when it is inconvenient. The goal is faster, safer models and product delivery, not a new bottleneck.
What You Will Do
Own Security Engineering Problems End to End
Take complex security problems from discovery and problem framing through design, implementation, deployment, adoption, validation, measurement, and iteration.
Use architecture analysis, threat modeling, code and configuration review, telemetry, adversarial testing, incident evidence, and data analysis to identify material risks and practical intervention points.
Build production-quality tools, services, integrations, controls, detections, automations, and reference implementations rather than stopping at findings or recommendations.
Work directly with product, research, platform, infrastructure, developer-experience, identity, data, and operations teams to implement changes in the systems they own.
Make tradeoffs explicit, deliver useful increments early, and improve the solution as operational evidence changes the understanding of the problem.
Illustrative first projects might include replacing a manual security control with automation, building detection for a class of identity misuse in training infrastructure, or hardening a developer workflow used by several teams. The starting point will follow the highest-value problem when you join.
Build Across Product and Operations Domains
Improve security in areas such as application and product security, cloud and platform security, identity and secrets, developer systems, software supply chain, data protection, security telemetry, detection and response, or incident readiness.
Turn repeated security needs into secure defaults, paved paths, reusable libraries, automated validation, self-service workflows, or platform capabilities.
Instrument systems and controls so their health, coverage, adoption, and failures are visible to accountable owners.
Support security incidents and exercises related to your areas of ownership. Investigate root causes, help contain and recover, and convert lessons into systemic improvements.
Contribute to technical standards, threat models, runbooks, architecture patterns, and documentation that allow other engineers to act without waiting for a security specialist.
When a team needs a design review, access exception, or risk decision, help find a workable path and stay connected to whether it holds up rather than acting as a pass/fail gate.
Operate as an AI-Enabled Engineer
Use models and agents to accelerate coding, query generation, investigation, analysis, testing, documentation, and workflow automation.
Validate AI-generated output using tests, representative data, source evidence, peer review, and direct inspection appropriate to the consequence of the decision.
Build AI-assisted security capabilities when they improve the system, with explicit authority, data access, evaluation, monitoring, human review, and failure handling.
Recognize when deterministic logic, conventional software, statistical methods, or human judgment are more appropriate than model-based reasoning.
Share effective AI-enabled engineering practices and help the team improve its speed without lowering its quality or security bar.
Contribute to a New Security Organization
Help establish the engineering practices, tools, operating mechanisms, and culture of a growing Microsoft AI security team.
Communicate progress, technical decisions, uncertainty, dependencies, and blocked work clearly. Escalate with evidence, options, and a specific decision request.
Seek expertise across organizational boundaries, mentor teammates where you have depth, and learn from principal and distinguished engineers.
Use central Microsoft capabilities when they solve the problem and build Microsoft AI-specific capabilities when the mission requires a different control point or operating model.
Raise risks and disagreements directly and respectfully, including when the easier answer is to stay quiet, and support execution once the accountable owner makes a decision.
What Makes This Role Different
The problem determines the domain. Demonstrated engineering depth and adaptability matter more than matching one narrow specialty.
You build rather than route work. The expected output is a changed system, adopted capability, or validated risk reduction.
AI is part of your normal toolchain. You should use it actively and responsibly while retaining engineering judgment and accountability.
You will help shape the team. The organization is early enough that strong engineers can influence its tools, standards, culture, and technical direction.
You work close to frontier-model development. The systems and workflows you protect support model training, evaluation, deployment, and operation at Microsoft scale.
What Success Looks Like
A material security problem has moved from ambiguity to an implemented and validated improvement.
A tool, control, detection, automation, secure default, or reusable pattern you built is operating in a real workflow.
Partner engineers can use the result without depending on you for every decision or operation.
AI-assisted work is faster while remaining testable, evidence-grounded, secure, and maintainable.
Incidents, near misses, and repeated friction lead to systemic improvements rather than recurring manual work.
The team has stronger engineering practices, documentation, and shared capability because of your contributions.
Required/Minimum Qualifications
Doctorate in Statistics, Mathematics, Computer Science, or related field
OR Master's Degree in Statistics, Mathematics, Computer Science, or related field AND 3+ years experience in software development lifecycle, large-scale computing, threat modeling, cyber security, anomaly detection, Security Operations Center (SOC) detection, threat analytics, security incident and event management (SIEM), information technology (IT), or operations incident response
OR Bachelor's Degree in Statistics, Mathematics, Computer Science, or related field AND 4+ years experience in software development lifecycle, large-scale computing, threat modeling, cyber security, anomaly detection, Security Operations Center (SOC) detection, threat analytics, security incident and event management (SIEM), information technology (IT), or operations incident response
OR equivalent experience.
Preferred/Additional Qualifications
Preferred qualifications are not independent pass/fail gates; candidates who meet some but not all are encouraged to apply.
Doctorate in Statistics, Mathematics, Computer Science, or related field AND 3+ years experience in software development lifecycle, large scale computing, threat modeling, cyber security, or anomaly detection
OR Master's Degree in Statistics, Mathematics, Computer Science, or related field AND 6+ years experience in software development lifecycle, large scale computing, threat modeling, cyber security, or anomaly detection
OR Bachelor's Degree in Statistics, Mathematics, Computer Science, or related field AND 8+ years experience in software development lifecycle, large scale computing, threat modeling, cyber security, or anomaly detection
OR equivalent experience.
CISSP CISA CISM SANS OSCP Security+
Strong software, infrastructure, or security-engineering experience, including writing, debugging, testing, and operating production code or automation.
Experience independently owning a complex technical problem from unclear requirements through implementation and sustained use.
Experience applying security judgment through one or more methods such as threat modeling, architecture review, code or configuration review, adversarial testing, detection engineering, incident investigation, or security-data analysis.
Experience directly implementing and validating changes in systems your team does not own, not only producing findings or requirements for someone else to build.
Ability to work through unfamiliar systems, identify a high-value intervention, and make progress without a predefined playbook.
Experience using telemetry, tests, operational data, or other evidence to make technical decisions and verify outcomes.
Clear written communication and the ability to explain technical work to technical and non-technical audiences.
Experience building security tools, platform capabilities, automated controls, detections, integrations, or response workflows that reached sustained production use.
Experience with incident response, threat hunting, adversary simulation, vulnerability research, product security, or infrastructure hardening.
Experience using large language models or agents in software engineering, security analysis, investigation, or operational automation with appropriate validation and oversight.
Experience contributing across more than one security or engineering domain.
Experience defining scope and priorities for problems that do not yet have an established solution or process.
Security Operations 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:
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.