About
I'm a solution architect and hands-on engineer. For twelve years I've built enterprise integrations and automation, usually from inside the customer's environment — from first discovery call through to the system running in production.
My current work is applied AI: multi-agent systems, MCP servers, RAG pipelines and LLM-driven workflows, built on Vertex AI and the Claude and OpenAI APIs. I write Python and TypeScript daily.
Work
Sep 2024 —
Present
Senior Solution Architect
Rossum — AI document processing for enterprise finance teams
- Led a six-month global rollout of AI invoice automation for a Fortune 500 semiconductor manufacturer across EMEA, APAC and the Americas, cutting per-invoice processing from minutes to under 60 seconds.
- Built Python/FastAPI microservices that move extraction output into Coupa, NetSuite and SAP over REST, OAuth2 and webhooks, with idempotent processing, retry logic and dead-letter queues.
- Designed three-way match validation, duplicate detection and tax rules across 15+ jurisdictions, reducing manual AP rework by around 40%.
- Work embedded with customer IT and finance teams from discovery through UAT and hypercare.
Oct 2016 —
Sep 2024
Solutions Architect → Director of Professional Services
Apify — web scraping and automation platform
- Built the global delivery and engineering organisation from scratch to 38 people; it accounted for over 70% of company revenue.
- Took project throughput from roughly 6 a month to 50+ by designing reusable solution templates and a marketplace connecting vetted freelance developers with enterprise customers.
- Designed and wrote Node.js/TypeScript crawlers and RPA automations for enterprise clients — pricing, catalogue and competitive data feeding analytics and decision-making.
- Owned departmental P&L, hiring and performance across four cross-functional teams.
Sep 2014 —
Jan 2018
Engineering Manager / Scrum Master
SmartEmailing — email marketing SaaS
- Ran the development lifecycle and Agile process, mentored product owners, and improved delivery predictability on a high-volume email platform.
Building
MCP servers
TypeScript · Vercel
Model Context Protocol servers for Claude and other LLM agents: an invoice management workflow server with auth, and a Nixtla TimeGPT forecasting server with Google Sheets integration, rolling backtests and anomaly detection.
Agent prototyping
Python · Vertex AI · Claude API
Agentic RAG and multi-agent pipelines for enterprise use cases, including prompt caching, batch processing and cost work.
Next.js · React · Node.js
A full-stack application — frontend and APIs — that I built and operate end to end.
Toolkit
- Languages
- Python (FastAPI), TypeScript / Node.js, React, Next.js, SQL
- AI
- Claude, OpenAI and Gemini APIs, Vertex AI, RAG pipelines, MCP servers, prompt engineering, deep reinforcement learning
- Data & infra
- PostgreSQL, MongoDB, Docker, GitHub Actions, AWS (Lambda, EC2), GCP, Hetzner
- Integration
- REST, OAuth2, webhooks, ERP systems (SAP, Coupa, NetSuite), web scraping at scale
- Delivery
- Enterprise accounts, team building, P&L ownership, Agile
Education
- 2014 — 2016
- Ing. (Master's equivalent), Information Systems and Technologies — Prague University of Economics and Business
- 2010 — 2013
- Bachelor's, Applied Informatics — Prague University of Economics and Business
Get in touch
If you're working on something I could help with — an ERP or API integration, an LLM agent, a pipeline that needs to hold up in production — I'm glad to take a look. Tell me what you're building and where it's stuck.
[email protected]
Email is the best way to reach me. You'll also find me on LinkedIn.