Journalist by training, Developer Advocate by practice. Ten years in cloud-native — seven in platform engineering and community organizing across Prague and Toronto, three running DevOps.Pink as an independent DA.

One of fewer than 250 Docker Captains worldwide. CNCF Ambassador. 8 vendor-recognized community titles across Docker, CNCF, AWS, IBM, Platform Engineering, Cypress, GitKraken, and Notion — earned through contribution, not credentials.

Founder of DevOps.Pink — cloud-native education for the agentic-AI era. 55K+ subscribers · 411K+ annual views · 130+ videos a year. Production-ready walkthroughs, demos, and code shipped within 48 hours of every major Docker release, validated against pre-release builds as a Docker Early-Access beta partner. Every framework I publish is tested in a live AI gateway I built and operate.

Docker Editorial on DevOps.Pink and the agentic-AI era

"Tatiana runs one of the fastest-growing independent cloud-native education practices in the industry, reaching tens of thousands of engineers worldwide. Her work focuses on making Docker, Kubernetes, and the emerging agentic-AI stack accessible through walkthroughs, demos, and production-ready code — with a particular emphasis on shipping close to release cycles so the community can learn alongside the newest features."

— Docker Editorial · docker.com/contributors/tatiana-mikhaleva

Selected publications, editorial bylines, and amplified releases

• Platform Engineering: The Empathy Gap: Why Your Platform Needs UX, Not Just APIs — authored byline; first publication of the Joy Metric, an NPS-style framework for Internal Developer Platform experience
• GitKraken Editorial: Two Developers, Two Continents, One Philosophy — International Women's Month feature profile spotlighting DevOps.Pink and the cross-border editorial workflow behind it
• Docker Official Channels: From Chaos to Calm: Docker Compose Saved My Dev Environment — DevOps.Pink short amplified across Docker's YouTube, X, and LinkedIn
• Docker Official Channels: My Dockerfile Was a Mess Until This — multi-stage build patterns, amplified across Docker's official YouTube and X feeds
• Beta Programs: Docker Early-Access partner across MCP Toolkit, Model Runner, and Sandboxes — friction logs shipped into GA features

225K+ combined views on the top 7

Platform engineer before Developer Advocate

Multi-year platform engineering advisory for a European bank across Kubernetes, CI/CD, Terraform, Vault, and private cloud — release-critical pipeline exposure in a regulated environment with audit-grade change controls. Built and ran Internal Developer Platforms hands-on for years before the DevRel pivot, which is why the platform engineering writing on DevOps.Pink reads like field notes rather than commentary.

Specialist degree in Journalism (Master's equivalent, Bologna Process) from Murmansk Humanities University — the unusual combination of journalism rigor, hands-on platform engineering practice, and measurable DevRel outcomes is what built DevOps.Pink into the editorial format it operates in today.

Every recommendation I publish is backed by production testing on real infrastructure that actually serves customer traffic in production — not theory, not vendor whitepapers, and not benchmarks rerun on a clean machine in a controlled lab where the real failure modes never fire.

Multi-tenant AI gateway · Claude · Codex · MCP

Architected and shipped a multi-tenant AI gateway with multi-LLM routing, per-tenant cost controls, and observability across Claude, Codex, and local models. Built using Claude Code and Codex as development partners, integrated with Model Context Protocol. The gateway is deployed as a live teaching platform in paid enterprise workshops, where engineering cohorts work hands-on with agentic AI infrastructure inside sandboxed tenants with full audit trails for everything each cohort builds.

My advisory work covers container platform selection, CI/CD architecture, Internal Developer Platform design, and agentic-AI integration patterns for engineering teams adopting LLM-powered tooling. DevRel typically means covering tools other people built — DevOps.Pink practice includes building production-grade tools first and then teaching from the operational scars and incident retrospectives that come with running them.

Speaking, workshops, and named partner programs

50+ live technical presentations across Toronto Cloud-Native & Container AI community and ambassador peer sessions across six vendor brand channels. Zoom sessions drew 300+ attendees during Prague organizing years; in-person meetups average 50+ engineers each. Enterprise workshops tailored to engineering cohorts: live coding, hands-on labs, architectural Q&A across Docker, Kubernetes, AWS, DevSecOps, and agentic AI stacks.

Independent practice funded through five named paid sponsors across container, security, observability, and DX categories: Proton (VPN and Pass), GitKraken, Linux Foundation, Plakar, and Rootly. Linux Foundation training partnership — code DEVOPSPINK for 30% off CKA, CKS, LFCS, and the full certification catalog.

Editorial voice, community building, and developer experience

Founded and run the Toronto Cloud-Native & Container AI community: zero members day one, scaled to 450+ active members in the first 12 months. Docker, Kubernetes, Terraform, and agentic-AI meetups. Vendor-sponsored demo nights. Full conference-format talks in the community's cadence. Before Toronto: Prague-based community organizing from 2016, bilingual teaching in Czech and English across two developer ecosystems.

Editorial voice built on opinion pieces, architectural post-mortems, and counter-narrative analysis from real production experience — titles like "A Nightmare on ELB Street," "Is Kubernetes Overkill?", and "Escape the YAML Prison" signal an editorial stance rather than a tutorial feed. Every script passes editorial review before the camera rolls; the journalism training shows in tighter lines, sharper hooks, and a higher bar for evidence than most DevRel channels currently meet on the cloud-native beat.

8 active vendor-recognized programs