I help CTOs and engineering teams turn AI-assisted coding into repeatable, reviewed work. We start with a real task, implement the workflow and give your team the checks and guidance to keep using it.
Work directly with me. Start with one team and one task. Scope, rate and an hour cap agreed before paid work.
AI workflows depend on the platform underneath. Identity, CI/CD, security and cost controls make a workflow maintainable. I connect AI adoption with the engineering systems your team already runs.
A focused implementation or regular hands-on support. You work directly with me, alongside a named owner in your team.
For teams introducing AI tools or getting inconsistent results from them.
We work through one task in your repository, improve the instructions and review process, and record what passes the agreed checks.
You keep: a worked example, repository guidance, acceptance checks and a handover.
Explore the AI engineering pilot →For teams ready to delegate a recurring engineering task.
Define the context, access and review boundaries for one workflow. Implement checks, bounded retries and a clear point for human intervention.
You keep: a reviewable workflow and recorded results to decide whether to extend it.
Discuss an agent workflow →For a specific AWS, cloud cost, CI/CD or migration problem.
Review the bottleneck, agree the change and work with your engineers to implement and verify it. Available as a standalone project or part-time support.
You keep: the implemented change, verification and maintenance notes.
Discuss a platform challenge →A separate four-week assessment of adoption, delivery bottlenecks, governance and measurement, from £18,000. Deliverables include a pilot design and a costed roadmap. It is optional; you can hire me directly for implementation or part-time support.
Illustrative assessment finding. A fictional example of the format, not a client result.
A completed platform migration and the capabilities built around AI-assisted delivery. These are programme-scale examples; your first engagement can start with a single team.
Azure DevOps was the starting platform. The migration moved more than 14,000 users to GitHub, establishing the platform foundation for the subsequent AI-SDLC work.
My enterprise platform work included this migration. It sits alongside my experience in cloud architecture, DevSecOps and developer enablement: moving a platform and helping engineering teams work with it.
I lead the AI-SDLC transformation, connecting Copilot adoption, governance, developer enablement, custom agents and measurement.
Adoption gained named owners and an AI Champions community. Engineering groups gained tailored sessions, reusable instructions and a controlled path for distributing agents.
Project summaries focus on my contribution and delivery outcomes. Further details or references are subject to client permission.
Practical methods for a new engagement, illustrated separately from the delivery experience above.
Explore how I measure AI adoption and put AI-assisted engineering into practice.
Selecting a video loads YouTube’s player. Videos do not start automatically. About video privacy
How I think about AI adoption beyond usage numbers, and what to look at next.
A walkthrough of my software-building workflow with Claude Code and a local Qwen model.
An introduction to GitHub Copilot HydraFusion and its setup.
Want to apply this in your team?
Discuss your project →Outcomes are measured against your own delivery data rather than vendor benchmarks.
The problem, success criteria and review boundaries, agreed with a named owner in your team.
A focused change on one team or one workflow, with planning and cycle time baselined before the work starts.
Delivery time, review effort and quality tracked before deciding what to expand. Autonomy is granted one workflow at a time, behind human approval gates.
Who this is for: CTOs, engineering leaders and platform owners with a concrete challenge, from a single team to an enterprise programme, with a named owner and agreed outcomes.
Consulting is delivered directly by me; larger programmes also need capacity from your own engineering team.
Yes. We can choose a first task and an approved tool, or improve a workflow using tools you already have. I work with your engineers on useful instructions, acceptance checks and review. An assessment is optional when priorities need a wider review.
Delivery metrics show outcomes. The AI value system measures licence activation, surface adoption, user-initiated versus agentic work, acceptance, planning or cycle-time change, quality and security, human intervention, and value per workflow.
It depends on the work. Architecture advice or an assessment can often start with delivery metadata, documentation and interviews. Implementation usually needs access to the relevant repository and tools. We agree the minimum access needed before starting.
You can review my professional background on LinkedIn. Further project details or references may be available, subject to client permission and confidentiality obligations.
Send a short project enquiry with the problem, your team and your preferred timing. We agree fit, scope, rate, a limit on hours, paid meeting time and availability before work starts. For a defined project, we also agree deliverables and acceptance criteria. The first step is to establish fit and clarify the work; sending an enquiry does not commit you to an engagement.
By not trying to reach every engineer personally. The transformation is principal-led and delivered through your own engineering organisation: I design the operating model, lead the first governed pilots, set the agent and review patterns, and train the team leads and champions who carry it into their teams. What scales is the system and the people you already employ, not consultant headcount. That is also why there is no delivery team to hand you off to.
No. You can start directly with part-time consulting or a scoped implementation. The four-week Transformation Assessment is a separate option for organisations that need an independent baseline and roadmap before a larger investment.
This is direct specialist consulting. If you need a fully outsourced delivery team or continuous operational cover, that requires a different staffing arrangement. We should agree the ownership and capacity your team can provide before taking on the work.
Tell me where you are stuck. You do not need a finished brief or a chosen package.
Prefer email? [email protected]
Remote consulting for UK, EU and US teams. Delivered directly by Matt Drankowski through Perun LTD.
Explore the frameworks at your own pace, or use the optional diagnostic to review how you measure AI adoption.
A 40-point self-assessment informed by NIST AI RMF 1.0, OWASP GenAI LLM Top 10 2026, ICO guidance, and practical delivery controls.
Run the assessment →Model activation, usage, and retained workflow value against the AI-SDLC measurement families, including the review-cost drag most ROI cases leave out.
Open the model →How to close the AI production gap: eliminating the ambiguity tax, handling context rot, and specifying work so agentic delivery stays reviewable.
Read the framework →
