I plan, run and report delivery end to end — Scrum and Agile cadence wired to AI-assisted engineering practice, so the roadmap stays true and the releases stay on time.
Requirements, estimates, risk — settled before the first sprint starts.
Backlog, sprint cadence, stakeholder reporting — run without the ceremony overhead.
AI-assisted estimation, review gates and reporting woven into every stage of the lifecycle.
Shipping, measuring, and feeding results straight back into the next sprint.
Product team, ~700 pull requests a month. Review was a steady source of queue time.
Introduced AI as a first-pass reviewer on every PR, so human reviewers start from a pre-filtered list of issues instead of a blank diff. Architectural decisions stay with humans.
Measured over three months.
Engineering team that needed a repeatable way to take a ticket to a reviewed pull request with AI, without losing control.
Led the design and rollout of a Claude Code workflow: ticket check, plan, build, independent review, ship.
On one complex ticket the run took 2 h 11 min and ≈ $80, added 81 tests, and the independent review caught 3 defects, including a latent bug the tests had been masking.
Standard-ticket figures are typical values. Complex tickets vary too much for an average, so one measured run is shown. About 15–20 tickets taken end to end so far.
Built and scaled cross-functional teams — BA, PM, Dev, QA — from 0 to 110 people across a portfolio of fintech and enterprise projects, without losing delivery quality along the way.
Designed a Discovery Phase framework for one engagement, then rolled it out across 7 client accounts — standardising how scope, risk and estimates get settled before a single sprint starts.
Built and ran the PM / PO / DM hiring pipeline for a fast-scaling delivery organisation — 50+ candidates assessed to date.
Independent Delivery Manager, sole proprietorship in Poland. From November 2021 I ran a portfolio of fintech and enterprise projects at Andersen Lab — banking, betting and media clients, end to end across Discovery and Delivery; before that, IT Project Manager in T&M and Fixed Price models. I started in December 2015 with a one-person IT venture, and that hands-on start is still how I work: Scrum and Kanban where they earn their place, AI woven into the SDLC where it actually saves time, never process for its own sake.
Education: Sukhoi State Technical University — Web Programmer.
Sole proprietorship in Poland — STDM, B2B engagements.
Claude 101, Claude Code in Action, Claude with the Anthropic API, Model Context Protocol, Agent Skills.
Scrum.org.
Professional Certificate, IBM · Coursera: 7 courses, started in 2023.
ICP-APO, ICAgile.
Andersen Lab: fintech & enterprise portfolio, Discovery to Delivery.
Cross-functional teams of 20–30 people, T&M and Fixed Price.
Discovery, build and support for 15+ SMB clients — end to end, alone.
Yes — banking and fintech portfolios are where most of the last four years have gone.
That's usually how it starts — most engagements begin mid-stream, not from a blank page.
A short call, on Telegram or wherever's easiest for you — no proposal before it.
