Skip to content

Practice areas

Custom software, AI and automation. Built to earn its place.

Three formats. The people who scope the work do the work. From one manual workflow to a full production platform. London-registered.

What we do

Eight things we do. One team.

  • 01

    Software architecture

    We design the system before we build it: what it does, how it fails, what it costs — and what should not be built at all.

    Explore
  • 02

    Workflow automation

    The repetitive work your team still does by hand. We map it, automate the parts that should be, and make sure nothing goes missing.

    Explore
  • 03

    Data & reporting

    One place where the numbers agree — ingest, warehouse and reporting your team can actually query and trust.

    Explore
  • 04

    Custom software

    Internal tools, client portals and products built around how you actually work — with AI only where it earns its place.

    Explore
  • 05

    AI in production

    Assistants, retrieval and models that stay reliable after launch. The MLOps plumbing: evaluation, monitoring, rollback.

    Explore
  • 06

    FinTech engineering

    Low-latency, regulated, real money. Prior to SerSan our CPTO was Senior Data Scientist at Revolut and VP, Quantitative Data Scientist at J.P. Morgan.

    Explore
  • 07

    Forecasting & risk

    Demand, pricing, signals, exposure. The mathematics behind decisions that are currently made on instinct.

    Explore
  • 08

    Fractional CTO

    A senior technical owner in the room for roadmap, architecture, hiring and vendor decisions, without hiring a full-time CTO.

    Explore

Engagement formats

Three formats. Start as small as you need.

  • 01 · Fixed scope

    Focused Diagnostic

    One workflow, product problem or system, looked at properly — or widened to a full technical audit when the whole picture matters.

    • The problem, framed in writing
    • Build, buy or leave it alone
    • Prioritised next steps
    • Effort and cost estimate
    • Optional: full technical audit

    2–6 business days

  • 02 · Build + ship

    Delivery Sprint

    Hands-on build in visible increments: design, implementation, testing, handover.

    • Design + implementation
    • Written acceptance criteria
    • Testing + QA
    • Handover docs + training

    2–8 weeks, depending on scope

  • 03 · Continuation

    Technical Partnership

    Continued development, support or fractional technical leadership — scoped separately each time.

    • Continued development
    • Maintenance + support
    • Optimisation + iteration
    • Fractional technical leadership
    • Roadmap + vendor decisions

    Scoped per phase, renewed on merit

How an engagement runs

Four phases. Sized to the job.

Start at any phase, stop after any phase. Price and scope are fixed before work begins. Process proportional to risk. Small projects stay small.

  1. Understand

    Duration
    One call
    Inputs
    The problem in your words, and what it costs you today.
    Outputs
    Agreement on the problem, and on what a good outcome is.
    What you get
    A written summary of what we heard and the options open.
    Risk reduced
    Buying a solution before the problem is clear.
    Decision point
    Stop here, or scope the smallest useful piece of work.
  2. Scope

    Duration
    2–6 business days
    Inputs
    Access to the workflow, system or data involved.
    Outputs
    The smallest valuable scope, priced, with acceptance criteria.
    What you get
    A written plan you can act on, with us or without us.
    Risk reduced
    Paying for a project nobody defined properly.
    Decision point
    Approve the scope, widen it to a full audit, or stop.
  3. Build

    Duration
    2–8 weeks, depending on scope
    Inputs
    Agreed scope, access, and one person who can decide.
    Outputs
    Working software in visible increments, tested against real use.
    What you get
    Code, accounts and documentation in your name. You own it.
    Risk reduced
    A system only the people who built it can run.
    Decision point
    Launch, extend the scope, or stop with what already works.
  4. Operate

    Duration
    Post-launch, scoped
    Inputs
    Real usage, the first weeks of incidents, user feedback.
    Outputs
    Measured results, fixes, tuning, an honest verdict.
    What you get
    Proof it did what it was supposed to do.
    Risk reduced
    A launch nobody measured and nobody owns.
    Decision point
    Hand over, or continue under a separately scoped agreement.

Most engagements run the full sequence. Many start at Scope or Buildbecause the problem was already clear. Some end with “don't build this,” and that's a successful outcome.

FAQ

The questions buyers actually ask.

It depends on the work. Diagnostics are fixed-price. Builds are scoped and priced per project. There's no minimum programme to buy into — one well-defined problem is a good place to start, and it can grow from there. Reply within one business day.

A Focused Diagnostic runs 2–6 business days. A Delivery Sprint runs 2–8 weeks, depending on scope. Longer builds are split into sprints you can stop between. We scope each one up front so nothing comes out of nowhere later.

Handover and a stabilisation period: documentation, architecture decision records, a walkthrough with whoever will run it, and us on hand while it settles. You own the code and the system outright — no licensing, no lock-in, no source held back. Continued development or support is available if you want it, scoped and priced separately.

Intake

Start a scoping conversation.

Four short steps. Two or three sentences is enough. A founder reads it and replies within one business day with what we'd build first.

Project brief

Step 01 / 04

Tell us about the work

What kind of work is this?
What are you trying to build, automate, or fix?

You don't need a spec or a budget. Just the problem you want gone.