tomáš machproduct engineer

ai-driven product engineer · prague

i build
the
whole
product.

I’m Tomáš. I turn ideas into software people use, whether it’s a production AI system or an open-source desktop app.

I do the whole stack myself: architecture, backend, frontend, AI agents, tests and deployment. Right now I work at Cleevio Labs and on a few products of my own.

13K+downloads of Na Pivo
#2free app in the Czech App Store
500+end-to-end tests on Uprate
1

real products with real users,
some mine, some built for businesses.

Each one started as an idea that somebody had to turn into working software. I like being the one who does that.

early alpha · open source

nuzky

Every video I made cost me two hours of captions, so I started building my own editor.

Nuzky is a free, open-source video editor for creators, with AI built into the editing itself. It renders on the GPU with its own engine written in Rust and runs speech recognition on your computer.

You cut a video by deleting words from its transcript, and captions are generated without uploading your footage anywhere. Over MCP, Claude Code or Codex can work directly on your timeline and take over the tedious parts.

(platforms)
macOS, Windows and Linux
(stack)
Rust, React, FFmpeg, local AI, MCP
(license)
Free and open source, GPLv3
nuzky.app

the raw take

umpauseretake

the cut

so I wanteda video editorthat cutsboring bits
top free · czech app store

#2

Na Pivo got there a few days after launch, between Google Gemini and foodora. I still find that ridiculous.

  1. 1Google GeminiProductivity
  2. 2Na PivoNavigationmine
  3. 3foodoraFood & Drink

na pivo

A stupidly simple idea that people downloaded more than 13,000 times.

It started as a joke: a compass that points to the nearest pub. I designed, built and shipped it alone, and the tiny side project grew into a proper app with beer tracking and friend groups. It’s still free and has no ads.

(platforms)
iOS and Android
(built)
Solo, design to backend
(stack)
React Native, Expo and Django
na-pivo.cz
cleevio labs · ai saas

uprate

A platform for running and growing mobile apps on the App Store and Google Play. Its AI agents help teams with reviews, store listings and pricing.

I’m the only engineer on it, so backend architecture, AI integrations, frontend, automated tests and releases are all my job. It has over 500 end-to-end tests. I also use AI agents to build, review and verify Uprate itself.

upratehq.com
cleevio labs · generative ai

uprate icons

An AI tool that generates app icons and App Store screenshots. I built nearly all of the engineering and turned the image generation workflows into a product that developers and app teams can use.

icons.upratehq.com

smaller
things

culinair

Culinair takes a recipe from TikTok, Instagram, YouTube or a blog and turns it into a proper one with ingredients, nutrition data and a shopping list. People pay for it.

culinair.app

vespra

An open-source AI companion for Discord, written in Go, with persistent memory, semantic search and tool calling. I built it to see how useful an assistant gets when it remembers more than one conversation.

github.com/tomasmach/vespra

eremvole

A client’s merch e-shop, moved from ageing PHP to Django and Next.js.

eremvole.cz
2

industrial hardware first,
ai software now.

Before I built AI products, I spent years writing software that talked to real machines. That changed how I think about reliability.

  1. 2021-2022

    dronpro

    Drone Engineer

    Enterprise DJI hardware like the Matrice 300 RTK and Phantom 4 RTK, with Zenmuse payloads.

    My first job where software and hardware had to work together.

  2. 2022-2025

    prusa research

    Software Developer

    Nearly three years at one of the best-known 3D printer makers: manufacturing automation, dosing-unit calibration over MQTT and Modbus, server-room access monitoring and internal tools.

    There I learned to write reliable, maintainable software and to understand the problem before writing code.

  3. 2026-now

    cleevio labs

    AI Developer

    I work on AI products and production SaaS. On Uprate I do all of the engineering and build its autonomous agent workflows.

3

ai agents write a lot of my code,
i decide what to build and check it.

I use agents because they let me build things that would otherwise need a much bigger team. They write code, review changes, run tests, inspect the running app and catch problems before users see them.

(ownership)

from idea to production.

I don’t need a perfectly written ticket to start. I can work out what needs to be built, make the technical decisions, implement it and get it to users.

(agents)

agents in my workflow.

I set up the systems that let agents take on more of the work, while quality, verification and technical decisions stay under my control.

(quality)

tested before it ships.

I care about automated tests, code review, solid architecture and the details users only notice when something breaks.

i like building things
that probably shouldn’t
be one-person projects.

So far that’s a pub-finding app thousands of people use, a whole video editor with its own rendering engine and AI products used by businesses.

I’m curious how far one engineer can take an idea with the right tools and a healthy obsession with making things work. I want what I build to end up in people’s hands.

tomáš mach
4

have an
ambitious idea?
let’s build it.

If you need someone who takes ownership, moves fast and sees the bigger picture beyond the code, we’ll probably get along.

I’m open to selected remote contract work, AI projects and product collaborations. Tell me what you’re building.