Hands-on ownership across technical direction, architecture, and cross-functional execution.
Governed Agentic AI Systems
I build the control planes that let agentic AI ship.
I'm Ziyad Mir, a hands-on principal engineer working at the boundary between model capability and production consequence: evals, permissions, review loops, audit logs, runtime orchestration, and rollback paths. At Roblox Ads/Discovery, that means turning AI into governed workflows for ads moderation, content understanding, and brand-safe creation surfaces.
Production workflows for agents, moderation, content understanding, and developer tooling.
Evals, policy gates, permissions, auditability, human escalation, and kill switches.
Experience across ads, discovery, high-scale infrastructure, reliability, and platform systems.
What I'm working on now
My current focus is the execution layer for AI systems that affect real users: the runtime, decisioning, evidence, and operations needed before agents can safely act.
- Ads moderation pipelines that convert creative, audio, destination, and policy evidence into auditable decisions.
- Runtime substrates for agents: sandboxed execution, tool access, artifact streams, task state, and rollback hooks.
- Reusable content-understanding systems for paid and organic surfaces: brand suitability, age appropriateness, discovery quality, and measurement.
How I work
The useful work is not writing strategy decks about AI. It is making the hard boundary explicit, then building enough of the system that other teams can safely depend on it.
Own the decision boundary
Define what the AI is allowed to know, decide, and do. Turn policy, safety, product quality, and business risk into explicit system contracts.
Build the reference path
Ship the first production workflow with APIs, queues, schemas, eval suites, operator tooling, observability, and human override loops.
Make autonomy earn scope
Start with controlled validation, measure disagreement and incident rates, then widen permissions only where evidence supports it.
Explore the work
Different ways to understand my approach and execution style.
Projects
Case studies across AI infrastructure, advertising systems, reliability, and developer productivity.
Writing
Technical essays on deployment control, agent runtimes, MCP, observability, content understanding, and engineering leverage.
Career overview
A concise overview of shipped systems, technical scope, and cross-functional engineering leadership.