Software consultancy / Warsaw / Worldwide
We help companies bring AI into their stack, redesign older backends that are slowing them down, and get their engineering teams working alongside coding agents. Founder-led, hands-on, no PowerPoint army.
What we build with
What we do
We don't sell platforms or shelfware. We design, build, and hand off working systems your team can run on Monday morning.
Plan the agentic workflow, set up RAG over your private data, ship custom MCPs for your tools. Designed for industries where every change has to clear security, privacy, and ops review, like finance, healthcare, and the public sector.
Pay down tech debt, redesign older backends, ship auth that holds up to a real review. OAuth2 / OIDC identity providers, Spring Boot rewrites, AWS infrastructure that won't wake you at 3 a.m. Pragmatic, not academic.
Onboard your devs to Claude Code, Copilot, and Codex with shared agent-standards, project skills, and MCP setups. Ship faster without losing quality. We pair-program, not just present slides.
How we work
Most engagements follow the same shape. We tell you what we'll deliver, by when, and what is out of scope.
One or two weeks. We talk to your team, read your code, map your constraints, and write up what we'd actually do. Fixed-fee, no commitment to continue.
Target architecture, ADRs, milestones, risks, and a proof-of-concept where it matters. You get a doc you could hand to any other team and they could execute.
We write the code, set up the infra, and ship in tight increments. Pair with your team or work alongside them. Always shippable, always reviewable.
Documentation, runbooks, knowledge transfer. Optional retainer for ongoing tuning, model upgrades, and the inevitable "one more thing" that comes up after launch.
Selected work
A subset, written up after delivery. Client names withheld where the contract requires it.
Designed and built an end-to-end identity provider, BFF layer, and AWS deployment topology for a new generation of frontend apps. From flow diagrams to Spring Boot code, Helm, and CloudFront / ALB configuration.
Architected a Spring Boot modulith with detachable services, Keycloak auth, and dual REST / SOAP integration into legacy government APIs. Flutter frontend, AI-augmented dev workflow throughout.
Open-source platform we use as a reference for client engagements. Per-request routing across LM Studio, OpenAI, Anthropic, Gemini, and MiniMax. Custom MCP servers for audio, web search, weather, and OCR. RAG pipeline backed by Qdrant.
More things we do
When AI integration or system architecture isn't quite the fit, these are the side practices we still take on. Each has its own page with detail and case studies.
Home Assistant systems, Zigbee infrastructure, custom addressable LED installations. From design to install, fully self-hosted, no cloud lock-in.
See smart home work →High-performance marketing sites, hosting and domain setup, SEO foundations. For founders who want to ship a real site, not a Squarespace template.
See web work →Common questions
No. We're strongest there, but we've shipped on GCP, Kubernetes, Python / FastAPI, and a fair amount of older code. The same engineering applies across most backends. The language and runtime change, the work itself usually doesn't.
If you're a 5-person startup, an AI readiness audit is probably the right shape. Big builds need more team. We'll tell you honestly when we're not the right fit.
Yes. We've shipped OAuth2 / OIDC identity providers, integrated with legacy SOAP government services, and designed RAG over private data for audited environments. Security is part of the build, not an afterthought.
We do those too, but they live on dedicated subpages so the homepage stays focused. Ask on the intro call.
Get in touch
Drop us a line about your project, your stack, and what you'd like help with. We read every message and reply within two business days. No slides, no pitch, no pressure.
