Skip to content
Flexiana
Menu
What we solve
How we work
Research
Work
Products
Company
Bring us a hard problem

Work

Work

Each case gives the problem underneath the request, the method, and what the evidence supports. Where a figure is illustrative rather than measured, it says so, because the alternative is a reader finding out.

4 cases is a thin record for ten years, and the reason is worth saying plainly. Everything we see is confidential, and we name a client only with written permission. So most of what we have done is not on this page, and the commercial outcomes that do exist belong to the clients who own them. Below sits what clients have let us publish, alongside 11 engagements we may name without quantifying. An owner who wants a reference gets one by phone, and what we require of ourselves before calling anything evidence is written down in how we validate.

Cases

Cases

Two systems running in production, one result carrying a machine-checked proof, and one worked example on synthetic data, labelled as such. Each card states its evidence status before it states its result.

Livesport coordinate One of the world's most-visited sports-results platforms runs on a fleet of internal systems that must stay correct and fast under enormous, spiky load, where a coordination or data error is visible to tens of millions of people within seconds. Flexiana designed, built and operated nine internal systems for Livesport in Clojure, ClojureScript and PostgreSQL, data-oriented services built to be correct under load and maintainable by the client's own team. In production Nine internal systems delivered and running in production behind a platform with more than 100 million users. Clojure / ClojureScript · Data-oriented systems · PostgreSQL Flexiana (own product. Gitrevio) predict Leaders could see engineering activity, commits, tickets, deploys, but could not connect that work to project cost, risk and margin until the accounting period had already closed. We designed and operate Gitrevio: a multi-tenant platform that ingests 31 engineering, project and HRIS sources, resolves contributors to canonical identities, and runs a decision-science layer that forecasts cost-to-complete and flags margin and delivery risk while intervention is still possible. In production A working, security-hardened product with 31 connectors and a 165-function decision-science layer, running inside Flexiana and offered commercially, the product and its pricing are public at gitrev.io. Bayesian forecasting · Identity resolution · ELT / Prefect Flexiana Decision Lab (Rainbow) verify For a decision rule that must never violate a hard constraint — a capacity, a contractual limit, an exclusivity — tests sample the input space but cannot close it. The residual risk is exactly the case the tests did not draw. We built Rainbow, a Clojure-native dependent-type theorem prover with a 1,200-line trusted kernel, and used the operations-research algorithm catalogue as its proving curriculum — machine-checking algorithms rather than merely testing them. Formally proved All 222 catalogued OR algorithms carry machine-checked proofs — zero pending, zero refuted; the prover checks signed rationals, finite vectors, Bellman/value-iteration and gradient methods natively. Dependent types · Proof-carrying certificates · Content-addressed proof data Illustrative example (not a client engagement) select A business with more demand than capacity must repeatedly decide which orders or projects to accept, at what price and start date — balancing expected margin against downside risk, scarce skills, subcontracting and delivery penalties. We model the choice as risk-adjusted selection under uncertainty: maximise expected contribution margin, penalised for downside (CVaR), delivery instability and client concentration, subject to capacity and skill-coverage constraints. Backtested On a synthetic benchmark, the risk-adjusted policy dominates a first-come / highest-price heuristic on the margin-versus-loss frontier — an illustration of the method. There is no client outcome in it. Stochastic optimisation · CVaR / risk-adjusted objectives · Constraint modelling

What each of these labels is allowed to mean: the evidence ladder

Selected engagements

Engagements we can name

Sports data at 100M+ users, fintech and payments, mortgage, govtech, SaaS, video infrastructure, fashion and HR technology. These are the clients who agreed to be named, with the scope we delivered. Where a specific commercial outcome is the client's to disclose, we hold it back and do not estimate it.

  • Information System · Software

    Livesport

    Built and still run nine internal systems behind one of the world's most-visited sports-results platforms, used by more than 100 million people.

  • Design · Software

    Stylitics

    Product design and software engineering collaboration with Stylitics.

  • Software

    Guaranteed Rate

    Software engineering engagement with Guaranteed Rate.

  • Software

    Cloudpermit

    Software engineering engagement with Cloudpermit.

  • Design · HRTech · Software

    Frankie

    HRTech product design and engineering.

  • HRTech

    Lifesize Video & Audio

    Video and audio infrastructure project.

  • FinTech · Mobile

    Zimpler

    Fintech and mobile engineering for Zimpler.

  • Blockchain · Design · Software

    Cicada Wallet

    Blockchain wallet engineering and product design.

  • Software

    Gitrevio

    Engineering intelligence platform — built and owned by Flexiana.

    Visit

  • SaaS · Software

    Crossbeam

    Co-development engineering on Crossbeam's partner-ecosystem platform, through product-market fit and into a period of rapid growth. Crossbeam's Director of Engineering tells the story himself, on video.

    Visit

  • FinTech · Software

    Starcity

    Extracted a payment microservice out of a monolithic application, including automating payouts from an external payment service to clients' bank accounts. Written up at the time.

Build stories

Build stories

Cards compress. These are the accounts we published at the time, written while the work was going on, including one where the client does the talking.

  1. 2022 What's co-development with Flexiana like? Crossbeam share their story Eric Nograles from Crossbeam shares how Flexiana engineers have been key partners during their product-market-fit phase and hyper-growth period. Ado Omerhodzic
  2. 2026 Building Data-Heavy Systems in Clojure Without Losing Simplicity As data volumes grow, complexity tends to grow with it, but Clojure offers a different approach using simple, data-centric architectures. Jiri Knesl
  3. 2023 The Creation of Frankie: A Service Design Tale How Flexiana used service design and iterative user testing to create Frankie, an HR recruitment tool that users love. Elaheh Nazari
  4. 2020 Building Payment Microservice in Starcity How Flexiana helped Starcity build a stable, scalable payment microservice by extracting functionality from a monolithic application. Pawel Zielonka
  5. 2020 E-commerce information system How Flexiana helped a major Czech e-commerce company rebuild their back office operations using microservices architecture. Mariusz Kotas
  6. 2020 Mobile app for Laclary instantly A 5-member Flexiana team developed a React Native mobile application for Laclary, an e-commerce platform with blockchain verification. Konstantin Mashirov, Roman Tosnar

More of the same: the engineering archive

Start with the decision

Bring us a decision

Describe the choice, constraint or operating problem. The reply says whether it fits an audit, a product, or a different route.

Bring us a hard problem See how Find, Prove, Build works