About

Boutique by design. Fixed-scope by principle.

The company

Thinking Machine d.o.o. is a strategic engineering advisory firm based in Zagreb, Croatia, founded in 2022. We work with compliance-sensitive enterprises and the vendors who sell to them — across cloud cost strategy, cyber resilience, and AI integration into existing systems.

We are deliberately small. The work is delivered by the founder personally, not staffed out to a delivery team. Each engagement is fixed-scope, time-boxed, and decision-grade. We take one to two new engagements per quarter.

Founder

Bruno Božić

Founder & Principal Advisor

I started writing software in the late 1990s and have been doing it professionally since 2003. Since then I have built and shipped systems in healthcare (hospital information systems, lab automation, e-Health certification), telecommunications, government R&D consortia, aviation, industrial IoT, and energy-sector procurement and NFR work. C#/.NET is my home stack; integration architecture is the thread that runs through all of it.

The organisations I have worked with — as employee, contractor, or fractional CTO — include national telecommunications operators, government agencies, healthcare networks, and EU-funded research consortia. I have managed teams of up to eleven across multiple European sites, mentored more than fifty engineers, and authored hundreds of pages of technical specifications and runbooks. I am currently Head of IoT at Reos GmbH (Hamburg), where I lead the engineering for a multi-vendor smart-building platform across hundreds of European sites. Thinking Machine is the channel through which I take selective external advisory engagements that do not conflict with that role.

I am also engaged in applied research on production systems for IoT device onboarding and runtime integration — multi-tenant certificate lifecycle management, secure mTLS at scale, and bridging affordance-centric semantics with the resource-centric REST APIs that increasingly dominate device-vendor clouds. The research informs the work, not the other way around.

Separately, recent applied work has covered real-time encrypted media architectures (WebRTC + SFU, end-to-end media encryption, self-hosted GPU inference for speech transcription and speaker diarization) for compliance-sensitive use cases where audio and video cannot leave the operator perimeter.

Recommendations from past colleagues across these roles are visible on LinkedIn. Attributable client-side testimonials are added to this page as engagements clear reference embargo.

Why fixed-scope

Most enterprise consulting engagements are designed to expand. Discovery becomes Phase 1; Phase 1 becomes Phase 2; the meter runs.

That model works for some problems. It does not work for the kind of problem Thinking Machine takes on, which is almost always: we need to make a decision, soon, and we need someone external to help us get there with rigor. Cloud migration go/no-go. NFR pre-contract due diligence. AI integration feasibility. These have a clear endpoint. The right deliverable is a document a board can read in fifteen minutes, backed by a model that survives a hostile read.

So we work in fixed-scope engagement shapes — Discovery Sprint, Rapid PoC, Fractional CTO retainer — with explicit out-of-scope items declared up front. If the engagement uncovers a bigger problem, we will tell you. We will not solve it through the same engagement.

What we won't do

We will not take work where AI, cloud, or compliance is not actually the right answer. We will not build the thing we recommend in the same engagement that recommended it (separation of advisory and implementation is structural, not stylistic). We will not sell you a partner relationship, a license, or a managed service. We do not have a sales team because we do not sell things — we sell time, capped.

If those constraints are a problem for your organisation, we are not the right fit, and we will help you find someone who is.

Thirty minutes, no slides, no pitch.

Book a 30-minute intro