Auto permission mode doesn't make Claude Code safe
Auto mode lowers the entry bar, but users stop learning the permission system and how to isolate their environment. Why manual mode, oddly enough, pushed people towards safety.
I help CEOs and CTOs of banks, fintech and product companies turn AI adoption into measurable delivery gains, by fixing the tools, the tech and the teams together. 20+ years hands-on, from writing Java to CTO at monobank’s marketplace.

Most companies are already through AI adoption: licences bought, tools rolled out, token dashboards glowing. Yet delivery hasn’t moved. The gap sits between the two phases.
Everyone uses the tools. Measured by breadth: who uses AI, how often, for what.
✓ done by most companiesThe business ships faster and better. Measured by outcomes: lead time, quality, cost of change.
→ where the value ismodels · agentic frameworks · harnesses · knowledge base
Agents are only as good as the environment and context you give them.
architecture · engineering practices · quality assurance · guardrails
Faster code generation hits the same bottlenecks: tangled architecture, slow pipelines, fragile tests. Without guardrails, risk reaches production faster too.
roles · skills · metrics · team & org design
If roles, skills and metrics stay the same, the speed-up disappears in handoffs and queues.
Most consultants work on one layer. I work on all three: two decades as an architect and CTO on the tech side, and as an org and agile coach on the team side.
A clear, realistic plan for getting delivery gains from AI, not just licences.
Hands-on implementation that moves from pilot teams to the whole engineering organisation.
Senior technical leadership without a full-time hire, or an interim CTO while you find one.
Practical, hands-on trainings that change how engineers work the next day.
Talks and workshops that give leadership teams a shared language for the AI shift.
Every engagement starts with a 30-minute intro call. No fixed price list: scope and format are shaped around your situation.
Book a call →500+ engineers
AI strategy consultant
Make AI a company-wide engineering capability rather than a collection of individual tools.
AI built into the way software is made: one SDLC, one agentic platform for the whole organisation, and shared knowledge that agents can use.
1,000+ engineers
AI adoption consultant and trainer
Roll out AI development tools across a large engineering organisation, and make sure people actually use them well.
200+ employees trained, and an internal AI champions group in place to keep adoption moving from the inside.
Ukraine’s largest neobank
Organisational & tech consultant, CTO advisor · 2023–2025
Grow organisational maturity and system stability across a fast-growing company, without losing its speed and culture.
A more mature organisation with new leaders and stronger engineering practices behind a product used by millions.
Also served as interim CTO of Market by Mono: OKRs for the whole product team, engineering scaled 6× in 6 months, LeSS across 6+ teams, features delivered to 9M+ users.
Builds digital banks in emerging markets
Organisational consultant, CTO advisor · 2022–2023
Scale product development across many teams while the architecture keeps evolving.
8+ teams working as one product organisation, on an evolving microservices architecture with a stronger release process.
One of Ukraine’s largest banks
Digital transformation consultant, CIO advisor · 2020–2022
Turn a large traditional bank into a modern engineering organisation.
A new engineering culture built around Continuous Delivery and collaboration, backed by trainings and workshops across the bank.

I’ve been building software since 2004: Java developer, tech lead, solution architect, delivery manager, CTO. Since 2017 I’ve worked as an independent consultant, helping banks, fintech and product companies with architecture, engineering practices and organisational change.
Today my focus is AI: helping engineering organisations move from using AI tools to actually delivering faster with them. I also teach engineers AI-assisted development and agentic engineering: more than 1,000 so far.
In 2009 I founded XP Injection, my training company. I also founded the JEEConf, XP Days Ukraine and Selenium Camp conferences.
Works in English · Ukrainian · Russian
Hands-on trainings for engineering teams through XP Injection, my training company since 2009: AI-assisted development, agentic engineering, engineering practices, technologies, architecture, Agile and more. More than 1,000 engineers trained in AI development so far.




Make Claude Code a full member of the team: from LLM basics and model management to context engineering, flexible configuration, MCP, hooks, skills and subagents, rolled out safely into team workflows and CI/CD.
Build AI agents that know your organisation and act on it: custom knowledge, tools and MCP, memory and context engineering, guardrails, agentic patterns in LangGraph, a custom harness and full observability.
Find and fix slowdowns, memory leaks and strange errors fast: logs, metrics and traces with OpenTelemetry, Prometheus, Grafana and Jaeger, plus thread and heap dump analysis on real-world cases.
200+ talks at conferences and meetups since 2008: AI and agentic engineering, architecture, Java, Spring Boot, engineering practices, Agile, QA processes, infrastructure and more.




WAWTech 2025
InterviewDOU 2021
Simplicity Day: Agile Magic 2020
Selenium Camp 2019 · JEEConf 2019 · Voxxed Days Minsk 2019
XP Days Ukraine 2019
XP Days Ukraine 2018 · DelEx 2019
Auto mode lowers the entry bar, but users stop learning the permission system and how to isolate their environment. Why manual mode, oddly enough, pushed people towards safety.
Token spend is a poor measure of value, but a good measure of adoption. Why an AI rollout needs two stages, each with its own metrics.
From the Agile Manifesto and digital transformations to 'we have nobody to configure JIRA'. How the Scrum Master role lost its meaning.
A 30-minute intro call: you tell me where you are, I tell you honestly whether and how I can help.