HI, I'M

MARTIN

FULL-STACK WEB DEVELOPER + AI AGENT ENGINEER

I build software where people and AI agents work together as teammates - agent workspaces, MCP integrations and agent-ready backends, shipped to production.

// 01

ABOUT ME

Full-stack developer with 5 years of experience building modern web applications and a master's degree in Computer Science. Strong background in both front-end and back-end development, with experience developing native mobile applications using React Native and VR applications in Unity. Passionate about technology, continuous learning, and building high-quality software, whether working independently or as part of a team.

Lately my focus is AI agents: not chatbots bolted onto a product, but agents that hold a real role in a team - with their own permissions, skills and tools, working on the same tasks and in the same threads as people. I build the infrastructure for that in QUESTPIE Autopilot, and I help companies bring agents into their own products and workflows.

STACK

REACTREACT NATIVENEXT.JSTYPESCRIPTPYTHONNESTJSPOSTGRESQLMONGODBStrapiDocker

AGENT STACK NEW

MCPLLM APIsTOOL CALLINGAGENT RUNTIMESpg-boss JOBSAGENT PERMISSIONSHUMAN-IN-THE-LOOP

EXPERIENCE

Full-stack DeveloperCODEUPP s.r.o.Feb 2026 - Jun 2026
Full-stack DeveloperQUESTPIE s.r.o.Jan 2022 - Present
Frontend DeveloperEliaš - IT Solutions s.r.o.Jul 2023 - Apr 2024

EDUCATION

MSc - Intelligent Software SystemsFIIT, Slovak University of Technology2023 - 2025
BSc - InformaticsFIIT, Slovak University of Technology2020 - 2023

LANGUAGES

SlovakC2
EnglishB2
// 02 - エージェント

AI AGENTS

Agents are only as useful as the system around them. I build that system: where agents get work, what they're allowed to touch, where they run, and how humans stay in control.

[ 01 ]

Agent workspaces

Places where people and agents share goals, tasks, knowledge and conversations - agents pick up work, ship it and report back in the threads the team already uses.

[ 02 ]

MCP servers & tools

Exposing your product, data and workflows to agents over the Model Context Protocol, so any capable agent can read, act and report through a clean, permissioned interface.

[ 03 ]

Agent runtimes

Running agents where it makes sense - a laptop, a VPS or managed compute - connected through a signed runtime, with background jobs and retries handled properly.

[ 04 ]

Agent-ready backends

One schema → database, REST API, typed client and admin. The same typed surface people use is the one agents use, so nothing has to be scraped or guessed.

[ 05 ]

Identity & permissions

Agents as first-class team members with their own roles, permissions and skills - never borrowed human accounts or shared API keys.

[ 06 ]

Agents in your product

Integrating agents into existing web and mobile apps - from a single assistant feature to a full multi-agent workflow behind the scenes.

01
Teammates, not tools

An agent has a role, an identity and an owner. You know who did what.

02
Everything visible

Every task, run and change request is inspectable by humans and by other agents.

03
Run it yourself

Your machines, your data, open source. No key to ask for, no lock-in.

04
Humans approve

Agents propose, people decide on what matters. Autonomy grows with trust.

GOAL
Team sets a goalin a shared workspace
TASK
Agent picks it upbased on role & skills
RUN
Works on a machinewith scoped tools via MCP
REVIEW
Opens a change requesthuman approves or redirects
SHIP
Ships & reports backin the same thread
// 03

SELECTED WORK

01
2024 - now / Open source · MIT

QUESTPIEAGENT-READY BACKEND

The ecosystem I build everything else on: a framework that turns one schema declaration into a backend, and Autopilot, which runs the company around it. QUESTPIE runs on QUESTPIE.

  • Postgres columns, REST API, typed client and admin, all generated from one collection
  • Built on Drizzle, Zod and Better Auth rather than its own runtime
  • You run it yourself. There is no key to ask for and no account to make
AGENT ANGLEA typed, schema-first API is exactly what agents need: predictable inputs and outputs, validated by Zod, guarded by real auth.
TypeScriptDrizzleHono · ElysiaZodBetter AuthPostgres · pg-boss
03
2024 - now / Technical partner

Nutrimeals

Nutrimeals is an EIT Food backed startup putting smart canteens and connected fridges into Slovak offices. I joined as technical partner and built the app people order from and the backend behind it. It runs multi-tenant, so Nutrimeals is one tenant of a gastro platform rather than a one-off app.

  • React Native app that maps the fridges and shows what is in them right now
  • Multi-tenant from the start, so a new canteen operator is configuration
  • Push notifications and offline support, because canteens sit in basements
QUESTPIEReact NativeExpoPostgresBun
04
2026 - now / Own product · pilot

Jubli

Every event gets a social space guests join by scanning a QR code. An event is a tree of spaces that changes as it moves from pre-event to live to the recap, so a wedding and a weekly pub quiz run on one engine with different presets.

  • Web, mobile and API in one monorepo, all of it on QUESTPIE
  • Space presets instead of code branches, so a new event format is configuration
QUESTPIETanStack StartReact NativeExpoHonoPostgresRedis
// 04 - OPEN FOR PROJECTS

WANT AI AGENTS ON YOUR TEAM?

Whether you want one agent feature in your app or a workspace where agents take real work off your team's plate - let's talk about what would actually move the needle.

Agent integrationBring agents into your existing web or mobile product.
MCP for your productMake your API and data usable by any capable agent, safely.
Autopilot setupRun your team with agents as teammates, on your own infrastructure.
// 05

CONTACT

LOCATION
Hnúšťa, Slovak Republic
AUTOPILOT
Talk to my agents too - they're on the team.

OFF THE CLOCK

GamingWarhammer paintingSci-fi fan - Star Wars / TrekMobile gamingCollector - figures & memorabiliaWorld of WarcraftLEGOMovies & TV series