Skip to content
willowark

Interactive software that trains people and holds attention

Willowark works the engineering side of interactive software: training and simulation games, the backend services a live product needs, telemetry pipelines that make player behavior legible, and build automation. We built a training video game with CCS for Jaguar Land Rover to help trainers perform better — applied games, where the goal is capability rather than sales.

It is ordinary software engineering held to game-shaped constraints. Backends stay stateless where they can and session-affine where they must, with account, entitlement, progression, and leaderboard services behind stable APIs so gameplay engineers are not maintaining infrastructure. Telemetry is designed as a schema first, then a pipeline. Build pipelines get treated as a product: cached artifacts, incremental cooks, parallel targets.

A first project with an interactive team is usually a single subsystem that is slowing everyone down: a build pipeline nobody has time to fix, a backend service written in a hurry during a milestone, or a telemetry pipeline that was never designed. For applied and training work, it usually starts with a prototype scoped to one competency and tested with real trainees. The constraint that shapes the engineering is that the engine and the content team set the pace, so our work has to fit around milestones and platform submissions rather than dictate them.

Reviewed

Illustrative: a software team's open office with monitors showing dashboards

Sound familiar?

If you've said any of these, we should talk.

Our build takes forty minutes and breaks constantly.

Build systems decay unless somebody owns them. We profile where the time actually goes, add artifact caching and incremental cooking, split platform targets across parallel agents, and put smoke tests on the output.

We need a backend, and our gameplay engineers are writing it.

That trade costs you the thing you are good at. We build the service layer — accounts, entitlements, progression, leaderboards, remote config, deployment — behind clean APIs, so your team stays in the engine.

We collect telemetry and nobody can answer a question with it.

Usually the events were never designed, just added. We define a schema tied to the questions design wants answered, build the ingestion and warehouse path, and give the team queryable sessions, funnels, and retention.

Our training content is a slide deck nobody remembers.

Interactive practice beats passive content when the skill is procedural, which is what we built with CCS for Jaguar Land Rover's trainers. We scope a simulation around one competency and instrument it so performance is reportable.

Our leadership wants to know if the training game actually works and we have nothing to show them.

That is a measurement design problem. We instrument the game around the competency it teaches — decisions made, errors, time to complete, retries — and report it alongside whatever assessment your learning team already uses, so the question becomes answerable with data from the sessions themselves rather than survey opinions.

How this industry actually runs

The operation as we usually find it.

Interactive teams live between an engine and a deadline. Unity and Unreal projects carry gigabytes of binary assets that source control handles badly, cook times that grow silently until a build is an hour, and platform requirements that arrive late. A live title adds a second discipline: accounts, entitlements, matchmaking and session allocation, save state, remote configuration, and analytics. Applied and training work differs again — the buyer is a learning group, content maps to a competency model, and results must be reportable.

Product, integrations, and the AI feature behind themAPIstructured callsYour productweb, mobile, APIBackend servicesqueues, jobs, evalsAI featureretrieval, review gatesData & integrationswarehouse, third parties

Product, integrations, and the AI feature behind them

Components:

  1. Your product (web, mobile, API)
  2. Backend services (queues, jobs, evals)
  3. AI feature (retrieval, review gates)
  4. Data & integrations (warehouse, third parties)

Connections:

  • Your product to Backend services (API)
  • Backend services to AI feature (structured calls)
  • Backend services to Data & integrations
Product, integrations, and the AI feature behind them

What we build

Starting projects that fit Gaming & Interactive.

  • Training and simulation games built around a defined competency model
  • Backend services for accounts, entitlements, progression, and leaderboards
  • Session allocation and matchmaking-style infrastructure for multiplayer products
  • Telemetry schema design, ingestion pipelines, and analytics for design teams
  • Build and release automation for Unity and Unreal projects across platforms
  • Remote configuration and feature-flag services so tuning and content ship without a client rebuild
  • Results reporting and LMS integration for training titles, exporting performance per competency
  • AI integration for in-game guidance, dialogue, and assessment with logging and guardrails

Capabilities we bring

Working in Gaming & Interactive?

Tell us the line.

What runs by hand, what is not connected, what you are trying to build. An engineer replies within one business day with whether and how we would approach it.

We reply within one business day. No newsletter unless you ask. Privacy

Common questions

What Gaming & Interactive teams ask first.

Are you a game studio?

No, and we would rather say so plainly. We are an engineering company that built a training video game and AI operating software for a large training organization, with CCS for Jaguar Land Rover. We take on applied games, simulation, tooling, and backend engineering.

Can you work inside our existing Unity or Unreal project?

Yes, and that is what most backend and tooling work looks like. We work in your repository, branching model, and coding standards, scoped to a bounded surface — a service, a pipeline, a tool — so the work stays out of gameplay code.

What does a training game project actually involve?

It starts with the competency, not the software: what should someone be able to do afterward that they cannot do now. We design the interaction loop, build a playable prototype fast enough to test with real trainees, then instrument it.

Do you handle platform certification and store submission?

We typically handle the engineering side and leave the account relationship with you. That means building against platform requirements from the start, automating the checks that can be automated into the build pipeline, and producing the technical evidence a submission needs. Platform holder accounts, agreements, and age ratings stay under your name, because they are commitments the studio makes rather than something a contractor should hold.

Can you take a prototype our team built and make it production-ready?

Usually yes, and the first step is an honest assessment of what to keep. Prototypes are typically right about the interaction and wrong about structure, so we look at what has to change for the backend, save system, and build to survive real users, then propose a path that preserves the parts that already work. Sometimes that is a rewrite of one layer; it is rarely a rewrite of everything.

Strategy. Software. Systems.

Engineering for Gaming & Interactive.

Describe the problem in your own words. An engineer reads it — not a sales script — and tells you plainly what it would take.