Principal Engineering Manager, Security Engineering
Microsoft AI · Mountain View, CA, Redmond, WA, New York, United States
Overview
We are looking for a principal people and engineering leader to build and lead security engineering for 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. You have built or substantially reshaped engineering-heavy security teams before and are prepared to do so again. Your job is to turn Microsoft AI security strategy into a team: hire and develop the engineers responsible for building controls, making security status legible enough to drive organization-wide prioritization decisions, and to build the enabling tools and systems that make the business run faster and more securely. You support this by running the roadmap, establishing the operating mechanisms, being the Directly Responsible Individual for some of security’s key metrics and deliverables, and remaining close enough to architecture, code, telemetry, and incidents to unblock execution directly.
Working with the Deputy CISO, Microsoft AI, you will set priorities, clarify ownership, manage performance, and establish the operating mechanisms needed for a new organization to deliver reliably. You will also join design reviews, pitch in on difficult problems, contribute code or prototypes when useful, participate in incidents, and help establish technical direction. You should know when to execute directly, when to coach, when to delegate, and when the team needs a deeper specialist.
Microsoft security teams earn trust by understanding the mission, staying close to the work, and remaining accountable for outcomes. We prefer useful progress and learning through implementation over perfect plans. We create clarity through explicit priorities, ownership, decisions, and measures. We communicate facts, analysis, uncertainty, and tradeoffs honestly.
What You Will Do
Build and Lead a Security Engineering Team
- Hire, onboard, coach, and retain a high-performing team of security engineers, architects, and operators with complementary depth across product and application security, cloud and platform security, identity, detection and response, security tooling, and other priorities as the organization evolves.
- Set clear objectives, provide actionable feedback, manage performance transparently, support career growth, and create meaningful opportunities for senior engineers to lead.
- Design the team's structure, roles, interfaces, and hiring sequence with the Deputy CISO. Identify where Microsoft AI needs dedicated ownership, where central Microsoft teams create leverage, and where a principal specialist is required.
- Create an engineering culture that values technical depth, operational ownership, clear writing, constructive disagreement, and durable systems over repeated one-off fixes.
- Raise the team's capability through technical review, mentorship, incident learning, reusable patterns, documentation, and deliberate investment in engineering quality.
Turn Strategy into an Executable Engineering Portfolio
- Translate security strategy, threat intelligence, incidents, architectural risk, partner commitments, and business priorities into a focused engineering roadmap with explicit outcomes, owners, milestones, dependencies, and measures.
- Run planning and execution mechanisms that make priorities, capacity, commitments, risks, and blocked decisions visible without turning engineers into status reporters.
- Work with program management to maintain a clear view of what the team owns, what partner teams own, and where a critical dependency lacks funded ownership. Escalate with evidence, options, and a specific decision request.
- Balance urgent risk reduction with investment in secure defaults, reference architectures, automated controls, detection, exposure analysis, identity systems, response tooling, and validation platforms.
- Use security and operational metrics to identify trends, prioritize high-impact work, expose gaps in measurement, and verify that controls and services are functioning as intended.
- Establish a sustainable operating rhythm for technical reviews, roadmap decisions, service health, incident readiness, corrective actions, hiring, and team development.
Lead Technical Work and Decision Support
- Engage deeply in architecture, threat models, design reviews, code, configuration, telemetry, and operational data. Help the team identify important attack paths, make sound tradeoffs, and choose practical implementation strategies.
- Contribute directly when it changes the outcome: build a prototype, review or write code, analyze an incident, query telemetry, test a control, automate a repeated task, or work beside an engineer through a difficult design.
- Lead ambiguous security engineering efforts from problem framing through implementation, adoption, validation, and measurement. Ensure work ends in a changed system or decision, not only a finding or recommendation.
- Help accountable owners in partner engineering teams make risk-informed decisions on proposed designs, exceptions, access changes, and launches. Offer practical alternatives and stay engaged through the outcome rather than issuing an approval or denial alone.
- Set a high bar for production engineering. Critical controls need clear ownership, safe deployment and rollback, observability, service-health expectations, incident procedures, and evidence that they reduce the intended risk.
- Use the tools, workflows, and AI-assisted security capabilities the team builds. If the team's systems are not useful or reliable enough for its own leaders to use, improve them before expecting broad adoption.
Protect Frontier-Model Development and Production
- Partner with research, model, platform, infrastructure, developer-experience, product, privacy, legal, and central security teams to protect the systems and workflows used for training, evaluation, experimentation, deployment, and model operations.
- Help create secure paths from research through repeatable production deployment, including practical controls for identities, secrets, data, workloads, administrative surfaces, software supply chain, shared infrastructure, and high-consequence operations.
- Improve security capability in partner teams through paved paths, reference implementations, reusable libraries, telemetry requirements, automated validation, and engineering support rather than review gates alone.
- Build readiness for incidents involving compromised identities, token or workload-identity misuse, data exfiltration, provider compromise, telemetry tampering, or compromise of model and training infrastructure.
- Integrate threat intelligence, hunting, adversary simulation, and incident evidence into detection priorities, engineering roadmaps, and security-control improvements.
- Turn incidents, near misses, repeated exceptions, and partner friction into engineering priorities and systemic improvements.
Operate Across Microsoft
- Build strong working relationships with Microsoft AI leaders, Microsoft CISO teams and platform owners before difficult launches, incidents, or prioritization decisions test them.
- Use Microsoft's existing capabilities when they solve the problem, improve or extend them when shared investment creates leverage, and build Microsoft AI-specific capabilities when the mission requires a different control point or operating model.
- Communicate technical risk, execution status, staffing needs, tradeoffs, and recommendations clearly. Distinguish measured facts, analysis, uncertainty, and recommendation.
- Represen