# Sr. Security Engineer, Operations — Microsoft AI

- Location: Mountain View, CA
- Posted pay: $120k–$235k/yr
- Posted: Sep 29, 2026
- Status: open (checked Nov 21, 2026)
- Apply (employer's site): https://job-boards.greenhouse.io/microsoftcorporation/jobs/4424233009
- Employer record: https://visajobs.io/company/microsoft.md

## Sponsorship history
- Tier: Strong history
- Occupation (SOC): Computer Systems Engineers/Architects (15-1299)
- Employer's certified LCAs for this occupation in CA, FY2024–FY2026: 239
- …in the whole US: 3,050
- With this exact title: 0
- Median DOL prevailing wage, level I / II: $82k / $115k

## Description
Overview
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, ment

> History, not a promise: past H-1B filings do not mean an employer will sponsor you or a given role. Data: U.S. DOL OFLC LCA disclosure (FY2024–FY2026 for role counts, certified LCAs) and the USCIS H-1B Employer Data Hub. Not legal advice. Methodology: https://visajobs.io/methodology
