About
I'm Dragan Lalos, an engineering leader based in Frankfurt. I build and scale teams and platforms in B2B environments-especially when growth exposes weaknesses in architecture, operations, and delivery. This page is a public snapshot of how I work: the principles I use, how I operate day-to-day, and the problems I'm typically brought in to solve.
Principles
Make outcomes measurable.
Define success with clear operational and delivery metrics (SLIs/SLOs, cycle time, incident impact) so engineering decisions stay grounded.
Design for ownership.
Clear boundaries, explicit decision records (ADRs), and end-to-end responsibility beat process-heavy coordination.
Stability enables speed.
Reliability practices aren't "extra work"-they remove drag and make product delivery predictable.
Reliability extends to AI.
Evaluation harnesses, citation enforcement, and failure-mode analysis aren't optional for production AI — they're the same reliability practices applied to a new substrate.
What I do
- Scale teams: hiring, structure, leadership layers, and clear expectations that support growth without chaos.
- Scale platforms: modernization, reliability engineering, and performance work so systems stay resilient under growth.
- De-risk delivery: operating cadence, metrics, and execution systems that make delivery predictable and aligned with goals.
How I operate
- Priorities: a tight quarterly focus with explicit trade-offs and clear owners.
- Metrics: delivery flow (lead time, deploy frequency) and reliability (SLIs/SLOs, MTTR) tied to outcomes.
- Rituals: lightweight planning, strong reviews, and consistent incident learning-not bureaucracy.
- Decisions: document the "why" (ADRs) and keep architecture aligned with product strategy.
Where I'm strongest
- Modernization programs (monolith to services, platform enablement, migrations)
- SLO-driven reliability and incident response
- Delivery systems (CI/CD, release engineering, quality gates)
- Stakeholder alignment and executive communication
AI engineering
I currently build AI systems in domains where "almost right" isn't good enough. As co-founder and CTO of VERAC AI Ltd, I work on AI decision-support for maritime and critical infrastructure operations. Separately, I build tools that help the Balkan diaspora in Germany navigate local bureaucracy — RAG-based systems with strict citation requirements and multilingual retrieval.
My approach to AI mirrors my approach to platform engineering: measurable outcomes, reliability practices, and rigorous evaluation before shipping. If the system can't be measured, it's not ready for production.
Values: clarity, ownership, measurable outcomes.
If you’re navigating growth pressure and want to sanity-check an operating model, share a short note with context. If there’s a real match, we’ll go deeper.