# Ondřej Úlehla

> Fullstack Developer & AI Engineer – I turn client problems into shipped software.

I'm a fullstack developer at NetGenium. I work across the whole lifecycle – sitting down with clients, analysing their needs, designing the solution, and then shipping it across frontend, backend and infrastructure. Over the past two years a growing part of that work has been bringing AI into our products and our own workflow.

Lately my focus is AI engineering: RAG pipelines over real client implementations, MCP servers, and automating processes that used to eat entire days – from customs clearance for the Czech Post to resolving client tickets, where the AI has the concrete codebase on hand and can find and fix the reported issue itself.

## Facts

- **Location:** Prague, CZ
- **Focus:** Fullstack · AI
- **Experience:** 5+ years
- **Website:** https://ulehla.dev
- **Email:** ondrejulehla@gmail.com
- **GitHub:** https://github.com/ondraulehla

## Experience

### Fullstack Developer & AI Engineer – NetGenium

*Jan 2021 – present*

https://www.netgenium.com

Enterprise applications on a low-code platform, client delivery, and integrating AI into the company’s products and internal workflow.

- Integrated AI across the platform end to end – from data and retrieval to the tools and interfaces used by developers and clients.
- Built RAG pipelines for the LLMs we use with clients, grounding answers in each client’s real implementation so the models reason over actual code and data.
- Automated internal processes with AI: generating applications in our low-code framework, and near-automatic client-ticket resolution – the AI has the concrete implementation on hand, so it locates the reported problem and fixes it.
- Automated the customs-clearance process for the Czech Post and built/customized their PostShop (postshop.cz) storefront.
- Built internal MCP servers and database-backed MCP servers for clients; currently building the new Hello.cz website together with a partner studio.

**Tech:** TypeScript, JavaScript, C#, SQL, React, Claude API, RAG, MCP

### Research Internship (exchange) – TU Wien

*Feb 2023 – Jun 2023*

https://www.tuwien.at

Research internship in Vienna during an exchange stay at TU Wien.

- International research experience alongside a demanding CS program.

### IT Administrator – Health centre

*Jan 2018 – Dec 2020*

IT administration and support during university studies.

## Education

### Ing. (Master’s), Software Engineering – Czech Technical University – FIT

*Sep 2019 – Jan 2025*

Included a research internship / exchange semester at TU Wien, Vienna.

## Skills

- **Languages:** TypeScript, JavaScript, Python, C#, C/C++, Java, SQL
- **Frontend:** React, Next.js, Vue, Astro, Tailwind CSS, HTML/CSS
- **AI Engineering:** AI agents & tool use, RAG pipelines, MCP servers, Claude API, Claude Agent SDK, OpenAI API, Sandboxed code execution, Prompt engineering
- **Backend & Data:** Node.js, FastAPI, .NET, PostgreSQL, Oracle, Redis
- **Infrastructure:** Docker, Kubernetes, Google Cloud, Git, CI/CD
- **Client work:** Requirements analysis, Solution design, Workshops & presentations, Technical writing

## Languages

- **Czech:** Native
- **English:** Fluent · C1
- **French:** Intermediate · B1
- **German:** Basic · A2

## Projects

- [MCPCost](https://ulehla.dev/projects/mcpcost.md) (2026) – A public board that probes every server in the official MCP registry and counts, exactly, what its tool definitions cost in Claude. 30 000 servers, a CLI and a GitHub Action for your own config, a JSON API and an MCP server.
- [codemode-lab](https://ulehla.dev/projects/codemode-lab.md) (2026) – MCP cost tools audit the tool definitions. That is the small number. What one tool call gives back is the number that decides whether your task runs at all, and code mode benchmarks do not measure it as its own term.
- [agent-flight-recorder](https://ulehla.dev/projects/agent-flight-recorder.md) (2026) – A flight recorder for AI agents – run a task N times in isolated E2B sandboxes, record every trace, measure pass rate, flakiness and cost per success, and diff the trajectories of two runs.
- [readme-ci](https://ulehla.dev/projects/readme-ci.md) (2026) – Run the code blocks in your README and fail CI when the quickstart breaks. Zero-dependency CLI + GitHub Action with Docker and E2B sandbox runners, published on npm.
- [mcp-sandbox](https://ulehla.dev/projects/mcp-sandbox.md) (2026) – Run any MCP server inside an isolated cloud sandbox instead of on your machine – one command, no code changes. Removes the blast radius of untrusted agent tooling.
- [agent-audit](https://ulehla.dev/projects/agent-audit.md) (2026) – An open-source security auditor for AI agent configurations – finds inline secrets, unpinned supply chains, blanket permissions and prompt-injection surfaces before they hurt you.
- [agent-lens](https://ulehla.dev/projects/agent-lens.md) (2026) – Drop an AI agent session transcript into the browser and see what actually happened – tool-call timeline, where the time went, tokens, cost and failures. 100% client-side.
- [Neural Networks in Virtual Reality](https://ulehla.dev/projects/neural-vr.md) (2025) – My master’s thesis: a web app that configures a neural network and a metaverse world that reads those parameters and builds a functional 3D model of the network – a “3D brain”.
- [AI Application Platform](https://ulehla.dev/projects/ai-framework.md) (2025) – A TypeScript monorepo platform for building AI-powered business applications – plugin architecture, typed APIs and a shared design system.
- [URL Shortener on Microservices](https://ulehla.dev/projects/url-shortener.md) (2026) – A production-style URL shortener built as a microservices system – cache-first redirects, stream-decoupled analytics and full observability, one command to run.
- [Trade-up Profit Calculator](https://ulehla.dev/projects/tradeup-calculator.md) (2024) – A data pipeline and calculator that finds profitable CS2 trade-up contracts – market data ingestion, probability model and a tested TypeScript core.

## Contact

Email ondrejulehla@gmail.com. Open to Forward Deployed Engineer, AI Engineer and fullstack roles.
