Available for B2B contracts · Gdynia & remote

Enterprise frontend, modernised without pausing delivery.

I'm Maciej Kankowski, an independent frontend tech lead. For 10+ years I've delivered enterprise platforms: technical direction, legacy modernisation and AI‑native workflows that change how a team ships.

Email me Download CV

Maciej Kankowski, smiling, in a light-blue polo shirt

Frontend chapter of 6in a 40+ engineer program

Release mastermonthly release train

React 19 · Vite · Node 22five apps migrated

10⁺years delivering enterprise web platforms

6frontend engineers under my direction

40⁺engineer program I work inside today

5sectors: HR, logistics, energy, finance, health

React
Node.js
TypeScript
Next.js
Vue.js
Vite
Playwright
Cypress
CI/CD
Multi-agent workflows
Claude Code
Design systems
Release engineering

Where I help

If one of these is your quarter, let's talk.

A product that aged while you shipped it

Every release costs more, and a rewrite is off the table.

A workstream without direction

Decisions happen inside pull requests; the same debate returns every sprint.

AI per person, not per team

Everyone prompts differently and the gain can't be measured.

Releases people are afraid of

Rare because risky, risky because rare. Fixing that loop pays first.

Current engagement

Fourteen months on one program: before and after.

An enterprise workforce-management platform for a global HR and staffing business. Five frontend apps, about 46 people, a monthly release train.

Frontend Tech Lead

Aug 2025 – present · under contract with Jit Team · end client under NDA

Active

Starting point

Four apps on Create React App, React 17–18, Node 16–18
Dependencies updated and releases assembled by hand
No shared component library, no UX process
AI used per person, with no shared rules

Now

All five apps on React 19, Vite, TypeScript and Node 22, migrated without pausing features
Renovate keeps dependencies current; a documented, feature-flag-first release process
A shared component library and a UX process: designed and approved before built
The team on Claude Code, policy first, reviewed by compliance and cybersecurity

Direction

A chapter of six engineers and a designer across four product teams, reorganised around the roadmap.

Release master

Six monthly trains and eight hotfixes for about 40 developers, with blameless post-mortems.

Redesign

Now leading the redesign of the entire app onto the component library, module by module.

Micro-frontend

Supported the team that rebuilt onboarding as one reusable form engine, changed in one place instead of per-app copies.

Safety nets

Security headers on every app, a frontend security audit, and automatic Zod validation of API responses.

Hands-on

Still writing production code: I built the core of the new planning module.

Fourteen months, in order

Aug–Oct 2025Joined; design-system groundwork
Nov–JanFeature flags, Release 5.0 across four apps
Feb–Mar 2026Library kick-off, Medallia, Claude Code
Apr–JunLibrary published, four-app migration live
JulGlobalLink Strings, new planning module
Aug–SepRenovate, Zod, UX process, team reorganised
NowLeading the redesign of the entire app

Engagements

Four ways to work together, each with an end date.

Fractional lead

Interim frontend tech lead

Set technical direction and unblock delivery.

Decisions written down
A team that decides without me

Modernisation

Legacy to modern React and Node

A phased migration, not a freeze.

Phases you can stop at
Delivery continues meanwhile

AI-native

AI-native team enablement

Workflows built into the pipeline, not into individual editors.

Working pipelines, not a deck
Guardrails, reviewable output

Audit

Frontend and delivery audit

Architecture, performance, testing and releases, reviewed.

Findings ranked by cost
Fixed price and duration

B2B through mackan.it Maciej Kankowski: my own equipment, no exclusivity, several clients in parallel. Engagement terms

A callWhat is slow, and what "done" looks like.

ScopeMilestones and acceptance criteria, approved first.

DeliveryIn your repo and process, reported against milestones.

HandoverDocs, patterns and pipelines stay with your team.

Earlier work

Five sectors, one way of working.

Also through Jit Team. End clients generalised by industry under NDA.

2022–2025Logistics · EMEA

Global logistics control tower

Cloud-native client for an EMEA stock-transfer platform with AI-assisted load recommendations. I owned end-to-end testing.

~40% fewer defects~€60k/mo unlocked~1.5h/day saved

2021–2022Health · Wearables

Digital health and wearables platform

Led a small team taking a web app and a native iOS app from MVP to production, owning iOS releases.

MVP → production

2020–2021Finance

Bond and debt-documentation platform

Modernised a security-critical system; as Scrum Master I guarded releases and drove performance.

3 new products seeded

2019–2020Energy

District-heating infrastructure platform

Greenfield platform with map-based 3D views. First engineer on the account, leading interactive UI.

~20 people at peak

Earlier still: native iOS (Swift), Android (Kotlin) and full-stack web.

AI-native engineering

Beyond autocomplete: AI as a reliable teammate.

01PlanA planning agent breaks the work down. A person approves the plan before anything is built.

02ImplementImplementation agents work inside your branching and CI, not in one person's editor.

03ReviewA review agent checks the change, then a person reviews it. Nothing AI-assisted ships unreviewed.

04TestTest agents run the change against the suite before it moves on.

05CI and releaseIt ships through your pipeline like any other change, with guardrails and documentation that stay with your team.

Planagent

Human approves

Implementagent

Reviewagent

Human reviews

Testagent

CI and releasepipeline

Multi-agent workflows

Agents hand work off; humans stay in the loop where it matters.

Wired into the SDLC

Generation, refactoring and review inside branching, CI and release.

Legacy, accelerated

Agents map, document and modernise a long-lived codebase.

Team enablement

Guardrails and review culture, covering the EU AI Act's AI-literacy duty (Art. 4).

Capabilities

What I bring, and where it ends.

Frontend

ReactNext.jsVue.jsTypeScriptViteVitestDesign systemsMicro-frontendsAccessibility

Leadership

Technical directionRelease managementArchitectureMentoringBusiness analysis

AI engineering

Multi-agent workflowsClaude CodeAgentic SDLCAI-usage policy

Backend and quality

Node.jsREST APIsGitHub ActionsGitLab CIPlaywrightCypress

Design and cloud

FigmaClaude DesignPrototypingAWS basicsDocker

Working knowledge of cloud, not deep infrastructure work.

Credentials

Gdańsk University of Technology · Gdynia Maritime University · Polish (native), English (B2)

The one-page CV

Kept current. The web version reads better on a phone.

PDF Read online

FAQ

Before the first call.

Who is Maciej Kankowski?

An independent frontend engineer and tech lead based in Gdynia, Poland, with 10+ years building enterprise web platforms. I run my own engineering practice, mackan.it Maciej Kankowski, and take fixed-scope B2B contracts covering technical direction, legacy modernisation and AI-native engineering workflows.

Are you available for new contracts?

Yes. I take on new contracts and run several clients in parallel, including the website clients of my studio. The fastest way to reach me is email.

How do you contract?

B2B, through my own registered business, mackan.it Maciej Kankowski, NIP 5882448371. Every engagement has a defined scope, agreed milestones and an end date. I supply my own equipment and software, take no exclusivity, and carry responsibility for the result of my work. Engagement terms.

Why B2B and not employment?

Under Polish law the test is how the work is actually done (Art. 22 of the Labour Code), not what the contract is called. My engagements are built on the factors that test looks at: a defined scope and an end date, payment for accepted results, my own equipment and software, my own choice of method, working time and place, no exclusivity, and liability for my own defects. The terms set it out point by point.

What technologies do you work with?

React, Next.js, Vue.js, TypeScript and Node.js, with CI/CD and Playwright or Cypress for end-to-end testing, plus design work in Figma and Claude Design. On the cloud side, simple AWS and Docker or Kubernetes configurations rather than deep infrastructure work.

What does AI-native engineering mean in practice?

Multi-agent workflows wired into the delivery pipeline rather than into individual editors: planning, implementation, review and test agents that hand work off, guardrails that keep the output reviewable, and documentation so the team keeps using them after I leave.

How does your work sit with the EU AI Act?

Engineering with AI tools is not a high-risk use under the AI Act, but it still carries duties. Client code and data go only into AI tools that meet the client's security and confidentiality requirements, and every AI-assisted change is reviewed by a person before it ships. Enablement covers the AI-literacy duty (Art. 4): written principles, a shared setup and training. If an engagement touches a high-risk area such as recruitment or worker management, its classification is settled in writing first. How I use AI.

Do you also build websites for small businesses?

Yes, through my studio at studio.mackan.pl: fast, mobile-first, SEO-ready business websites in Polish or English, handled end to end in fixed-scope packages.