Heikelä

Independent technology advisory · London and the Nordics

Technology judgement that provides an external perspective for the decisions that matter most, and helps you avoid costly blind spots.

I help investors, founders and engineering leaders see a software organisation clearly: what’s solid, what’s fragile, and what the next stage of growth will demand of its code, its architecture and its people.

Book a 30-minute call See the three services

Services

Fixed scope engagement models for three pain points

Each has a defined deliverable, a fixed price and a set turnaround. You buy an outcome, not a block of days.

For funds, acquirers and founders before a raise

Technical due diligence

An independent read of the technology, the team and the risks, written for an investment committee.

  • Architecture and codebase, mapped against what matters to the deal
  • Team, delivery process and AI-tooling maturity
  • AI-era risks older frameworks miss: comprehension debt, duplicated AI-generated code, model and vendor dependencies
  • Findings translated into cost and timeline, with a remediation plan

Rapid screen: red-flag read within 72 hours of data-room access. £3k (1)

Full tech DD: report in 10 working days.

DD readiness: for founders, fix the story and the risks before investors find them.

For CTOs and heads of engineering

AI-era engineering enablement

For the team whose board keeps asking "what's our AI plan?" or "do we still need X if AI can do it for us?". Practical change in how the team builds, from someone who builds this way every week.

  • A maturity benchmark of how the team builds with AI today
  • A sequence of efficiency improvements and experiments the team chooses and owns
  • Hands-on coaching in the tools and workflows, not slideware
  • Measured quarterly, improved weekly

AI operating-model quick check: high priority actions and current state compared with evolving industry practice, delivered within 6 working days. £2.5k

Full AI-in-product-development diagnostic: four weeks, with a board-ready readout and a 12-month roadmap.

Enablement programme: multi-week, run alongside the team's normal delivery.

Workshops: half-day or full-day sessions for engineering leaders and teams.

For boards and CTOs of European companies

Dependency and cost review

Is what you pay for cloud and AI aligned with your goals. Which of your vendors present a data sovereignty issue or problematic lock-in, and what would it take to change that.

  • Exposure mapped across cloud, AI models, SaaS and data flows
  • Exit feasibility and cost for the dependencies that matter
  • Cloud and AI spend: where it goes and what to cut
  • A board-ready posture report with a prioritised plan

Review: about two weeks, fixed price. [Price]

Cost review only: a shorter, savings-focused version. [Price]

Concrete risk reduction a board can act on.

Method

Concerns, choices, code

Every engagement uses the same model of a software organisation. What it cares about only reaches its code through the choices it has made. The gaps between the three layers are where risk lives, and where a deal, a raise or a rebuild gets expensive.

  1. Outer layer

    Concerns

    What the team, and its investors or acquirers, care about: security, auditability, cost, speed, compliance.

  2. Middle layer

    Choices

    The architectural commitments made to address them, each somewhere between exploring, proposed, committed, in progress, realised and superseded.

  3. Inner layer

    Code

    What actually exists. AI now makes this layer cheap to map, and makes it possible to read the unwritten choices back out of it.

Six gaps we look for

  • Unaddressed concernSomething that matters, with no choice made to handle it.
  • Orphan choiceA commitment nobody can trace back to a concern. Worse if it works against upcoming concerns.
  • Stalled choiceCommitted, started, never finished. Half-migrated isn't half-done.
  • Contradicted choiceCode that works against a decision the team believes it made.
  • Entangled choiceChoices so bound together that none can change alone.
  • Implicit choiceA pattern in the code that was never actually decided.

About

Mikko Heikelä

CTO, consultant and still a builder. I ship production software with AI agents every week, so the advice comes from practice.

Seven years as CTO of eduMeA London edtech company. Scaled engineering from the pre Series A team through Series B, then led the move to AI-leveraged delivery and a smaller team in pursuit of profitability. Still at eduMe much of my week
Thirteen years at FuturiceR&D-heavy consultancy work in Helsinki, Berlin and London, for Nokia Research Center and Nordic and UK brands, combining agile engineering, human-centred design and lean startup.
Both sides of the tableI have prepared a company for investor scrutiny and know what diligence looks for. Done DD on a product we acquired and integrated into eduMe. These inform what I look for when I'm the one reviewing.
Nordic by trainingPhysics at Helsinki University of Technology. Finnish, based in Kent, working across London and the Nordics.

How engagements work

Fixed scope, fixed priceAgreed up front. Anything outside scope is quoted separately before it starts.
Deliverables, not daysYou get a named output on a set date, written for the people who have to act on it.
Independent and confidentialNDAs both ways. Conflicts are checked before any engagement begins.
A few clients at a timeYou work with me directly, not with a junior team.

Taking on founding clients for Q4 2026

A deal coming up, a raise to prepare for, or a team to move faster? A 30-minute call is enough to tell whether I can help.

mikko.heikela@pm.me LinkedIn