Skip to content
Khaled Zaky

Hey, I'm Khaled

I build platforms for governing and evaluating AI, responsibly and at scale.

Head of AI Governance & Evaluations Platform Engineering · RBC Borealis

I've spent the last 12+ years designing and operating large-scale cloud, identity, and platform systems, mostly in highly regulated environments where security isn't optional and the stakes are real.

Right now, I lead AI Governance & Evaluations Platform Engineering at RBC Borealis, within RBC's AI Group. My work sits between enterprise AI governance and hands-on platform engineering: building the evaluation systems, scorer libraries, datasets, release gates, runtime monitoring, and control automation needed to govern, evaluate, and deploy AI responsibly at bank scale.

That means working across governance, audit, risk, second line, regulators, and business teams — but also getting deep into the engineering machinery behind model and agent evaluations. I think governance should not be a manual checkpoint at the end. It should be designed into the platform from the start.

Before stepping into this role, I built and led the bank's agentic AI platform at RBC Borealis, where I designed the orchestration layer, agent runtime infrastructure, identity integration, and the observability systems that make AI workloads debuggable in production. That engineering foundation is what shapes how I think about governance today.

Before this, I was a Senior Product Manager (Technical) at Amazon Web Services, where I owned product strategy for AWS sign-in security, MFA, passkeys, and credential risk. I led the design and launch of passkeys and multiple MFA devices for AWS IAM, and played a key role in root MFA enforcement, the kind of work that touches millions of customers and has zero room for error. I also served as Amazon's representative on the FIDO Alliance board and contributed to the W3C WebAuthn Working Group, where I helped shape global standards for passwordless authentication.

Before AWS, I was Director of Product Management for Cloud & Transformation at RBC, where I owned strategy and roadmap for the bank's enterprise cloud platform and led cloud, infrastructure, and AIOps product teams. Prior to that, I held product and architecture roles at TD Bank and Rogers Communications, and started my career writing code at BlackBerry, Siemens, and McMaster University.

I also teach Product Management and AI at BrainStation, where I mentor future product leaders and help them connect technical depth with real-world product thinking.

What drives me is a pretty simple idea: intelligent and secure systems should be accessible to everyone. I care about making things that are usable, trustworthy, and built to last.

Background & Credentials

BEng, Software Engineering

McMaster University, 2014

P.EngAWS Certified ArchitectAWS Certified Cloud PractitionerCIDPRO FoundationCertified Scrum Product OwnerFIDO AllianceW3C WebAuthn

Current focus areas

AI governance and evaluations
Identity and authentication
Cloud and distributed systems
Product strategy and platform design
Usable security and developer experience

How I work

  • Start with customer and operational reality
  • Treat security and usability as design constraints, not tradeoffs
  • Prefer clear platform primitives over one-off solutions
  • Write to clarify decisions and align teams
  • Build iteratively with strong feedback loops

About this site

  • A living archive of what I'm learning and building
  • Notes and essays on platform design, identity, and AI
  • A sandbox to practice modern tools and publish in public

Technical details

  • Designed and coded on a Mac
  • Built with Astro and Tailwind CSS
  • Coded in Windsurf (now Devin) and Claude Code
  • Source code hosted on GitHub
  • Deployed with AWS CodeBuild on each push to GitHub, with local caching for faster builds
  • Hosted on Amazon S3 with Origin Access Control (public access fully blocked)
  • Served through Amazon CloudFront with HTTPS enforcement, gzip/Brotli compression, and security headers
  • URL rewriting handled by a CloudFront Function at the edge
  • DNS managed by Amazon Route 53, secured with a wildcard TLS certificate from ACM
  • Blog posts are enriched and polished by an AI agent powered by Amazon Bedrock: I write the draft, the agent does research and styling using my own voice profile
  • Data charts are code-rendered as SVG (no AI-generated images, no stock photos)