Karel Mareš

AI-native full-stack product engineer Czech Republic · EU

AI-native products, built with senior engineering judgment.

I help founders and product teams turn complex problems into reliable software—combining 10+ years of full-stack engineering with AI-assisted research, delivery, and automation.

See selected workDiscuss a project
Available for selected product and consulting work
Scroll to enter
AI product strategyAgentic developmentWorkflow automationFull-stack systemsProduct engineering

Selected systems,
built for real use.

Different domains. Same standard: understand the real constraint, build the right system, and leave it maintainable.

01

Geospatial platforms · Operations · AI documentation

AVAG

At AVAG, I build and maintain three connected products: a 2D/3D geospatial client, an internal operations backoffice, and an AI documentation system that turns validated product knowledge into answers people can trust.

  • Used in 80+ countries
  • Svelte · Three.js · OpenLayers
  • Go · Templ · HTMX
  • TypeScript · Astro · AI retrieval
  1. Map client

    Responsive 2D/3D workflows for imagery, point clouds, measurements, and technical data.

  2. Operations backoffice

    Permission-aware support and business workflows, with customer, billing, and activity context in one workspace.

  3. AI documentation

    Validated content, retrieval, and an embedded assistant designed to answer from product knowledge—not guess.

Visit AVAG
02

E-commerce · Long-term frontend ownership

Libristo

For nearly a decade, I have owned and evolved the customer-facing frontend of Libristo’s large Laravel application. It supports browsing and buying across 22 million titles in more than 100 languages.

  • ~10 years ownership
  • Laravel · PHP · JavaScript
  • 22M titles · 100+ languages
  1. Long-term ownership

    Nearly a decade of shipping improvements while preserving a mature, business-critical storefront.

  2. Laravel application

    Frontend work inside the core Laravel application—not a detached presentation layer—so every change must fit existing product and backend constraints.

  3. Commerce at scale

    Keep discovery and checkout clear across 22 million titles and more than 100 languages.

Visit Libristo

AI is the multiplier.
Engineering is the foundation.

I use AI to research options, challenge assumptions, prototype, write and review code, automate repeatable work, and test earlier. You get useful feedback sooner; I remain accountable for architecture, security, trade-offs, and quality.

AI-assisted delivery

Specialized agents help research, prototype, implement, test, and document in parallel. You see useful iterations sooner; I review and own every decision.

AI product features

I turn models into dependable product behavior with tool use, retrieval, structured outputs, evaluations, guardrails, and clear fallbacks.

Workflow automation

I connect APIs, data, and n8n workflows to remove repetitive work and shorten handoffs—without hiding critical steps from your team.

Full-stack foundation

Ten-plus years across interfaces, APIs, databases, and deployment helps me fit AI into systems that stay fast, understandable, and maintainable.

AI changed how fast software can be built. It did not remove the need for judgment.

Faster iteration. Clear ownership. Software your team can keep building on.

AI helps me inspect more approaches, surface risks earlier, and spend less time on repetitive work. It does not decide what matters.

I stay responsible for architecture, trade-offs, security boundaries, and quality. Clients get faster feedback, fewer dead ends, and software their team can understand, operate, and extend.

See my experience on LinkedIn
10+years
shipping
Agentic AIn8nTypeScriptPythonReactTanStackNode.jsPostgreSQL
Agentic AIn8nTypeScriptPythonReactTanStackNode.jsPostgreSQLSvelteLaravelGoHTMX

Have a hard product problem?

Let’s find the
right way through.

Discuss your projectYou’ll speak directly with me—no sales handoff.