Loading page…
—Approach & method
The build process reflects senior delivery habits, not portfolio theatrics.
How a workflow becomes software here: mapped, built in slices, verified with numbers, and handed over with material another team can actually use.
TypeScript✦React✦Next.js✦Node.js✦Python✦PostgreSQL✦Prisma✦OpenAI APIs✦Vercel AI SDK✦AWS✦Azure✦Docker✦Redis✦GraphQL✦NestJS✦Tailwind CSS✦LangChain✦MongoDB✦TypeScript✦React✦Next.js✦Node.js✦Python✦PostgreSQL✦Prisma✦OpenAI APIs✦Vercel AI SDK✦AWS✦Azure✦Docker✦Redis✦GraphQL✦NestJS✦Tailwind CSS✦LangChain✦MongoDB✦
AI SaaS·Enterprise systems·E-commerce·Voice agents·Tender intelligence·Automation·Performance rescue·Design systems·Cloud & DevOps·AI SaaS·Enterprise systems·E-commerce·Voice agents·Tender intelligence·Automation·Performance rescue·Design systems·Cloud & DevOps·
01Operating method
Five steps, each one leaving something checkable behind.
The CV spans architecture, implementation, performance, DevOps, automation, and user-facing systems. The method keeps those responsibilities connected instead of handing them across a wall.
01
Step 01
Map the operating problem
Before architecture, the workflow. What enters the system, who touches it, where it stalls, and what evidence proves it worked. Constraints become requirements.
Stakeholder interviews
Workflow audit
Risk & bottleneck log
Success metrics agreed
Engagement shape
Discovery slice → build slices → verification → handoff. Priced per slice, with a definition of done agreed before the first commit.
Fixed-scope slicesWeekly demosPreview deploysDocumented exit
02Principles
Six rules that decide what gets built and what doesn't.
01
Performance is a design input
Budgets for load time, bundle size, and cost-per-request are decided at the start, not negotiated after launch.
02
Evidence over adjectives
Screenshots, metrics, repos, and live URLs. If it cannot be shown, it does not go in the case study.
03
AI needs boundaries
Models get clear scopes, structured outputs, citations, and escalation paths. Confidence is designed, not assumed.
04
Small surface, deep quality
Fewer features, properly finished: accessible, tested, documented, and fast under real content volume.
05
The operator is the user
Dashboards are built for the person doing the job at 4pm with thirty tabs open, not for the demo room.
06
Handoff is part of the work
A system nobody else can maintain is unfinished, no matter how well it performs this quarter.
03Field reports
What it is like to have me inside the build.
References available on request, including contacts at the companies behind the work below.
“He took over a build that had stalled for two quarters, shipped the missing 30% in six weeks, and left documentation the internal team still uses a year later.”
Engineering Lead
UK e-commerce group
“The tender analysis went from three days of reading to an afternoon of deciding. The eligibility scoring alone paid for the project.”
Bid Director
Public-sector contractor
“Rare to find someone who argues about query plans in the morning and about button hierarchy in the afternoon — and is right about both.”
Product Founder
Voice SaaS startup
04Working agreements
The unglamorous parts, written down in advance.
Most engagement friction comes from unstated assumptions. These are stated.
Communication
Written async by default, one live call per week, demos every Friday.
Code ownership
Your repo, your infrastructure, your accounts. No hostage situations.
Quality bar
Typed end to end, tested at the seams, reviewed before merge, documented on exit.
Change control
Scope changes get a written delta with effort and impact — never a surprise invoice.
AI usage
Assisted, reviewed, and accountable. Generated code is read before it ships.
Availability
CET business hours with overlap into UK mornings and Swedish afternoons.
05Questions
Answered before
you have to ask.
Something not covered? Send it over — a direct answer costs nothing and saves a call.
Ask a questionWith a paid or scoped discovery: the workflow, the constraints, the expected business outcome. You get a practical engineering path and a fixed first slice, not a vague estimate.
Yes — often as the senior builder inside a smaller team, or as the person who rescues a stalled build and hands it back documented. Remote, overlapping CET hours, comfortable with UK and Swedish schedules.
Most AI SaaS slices land between €6k and €25k depending on integration depth, document volume, and compliance needs. I quote per slice with a clear definition of done.
Frequently. The first week is a code health pass: architecture map, risk log, performance baseline, and a prioritised list of what to fix, what to keep, and what to rewrite.
Runbooks, observability, and a recorded walkthrough. Optional retainer for iteration, but dependency on me is treated as a failure mode.
10Get in touch
Bring the product pressure,
the constraints, and the
expected business outcome.
Send the desired outcome, the users, the current bottleneck, the stack, and the timeline. You'll get a practical senior engineering path for the build — not a brochure.