building & productionizing AI — from traditional ML to GenAI & LLMs
Contact Martin
I love bringing AI into production. My passion is turning data and modern AI into solutions that actually ship and create value. But, when I am not building software or AI systems, I enjoy hiking, sports and playing games with friends.
My area of focus is the end-to-end AI / ML lifecycle — from data preparation and model development to deployment, monitoring and continuous improvement — spanning traditional machine learning and modern GenAI / LLM applications, with particular focus on the insurance and e-commerce domains. I create value by understanding the business side of problems, delivering a return-on-investment, and combining strong, client-oriented communication with mentoring and technical leadership.
About MeAn architecture case study on Retrieval-Augmented Generation: how I design a pipeline that answers from a company's own documents — with citations, anti-hallucination guardrails and an evaluation loop. Covers ingestion, chunking, embeddings, vector search, re-ranking and grounded generation, with annotated LangChain-style code.
See Case Study
How I structure multi-step LLM applications as an explicit, stateful graph — with nodes, conditional edges, tool calls, bounded retry loops, checkpointing and human-in-the-loop. A walkthrough of a self-correcting document assistant, with annotated LangGraph-style code and the key trade-offs.
See Case StudyDesigning tool-using, multi-agent solutions: a planner/executor pattern with specialised agents, tool & MCP integration, memory and a critic that verifies before finalising — plus how to keep agents bounded, observable and properly evaluated. Honest, PoC-level architecture thinking.
See Case Study
This interactive dashboard was created in collaboration with The Prague University of Economics and Business in order to help stakeholders understand potential effects of social media posting and behaviour and its correlation to selected features of 2020 Senate elections in Czechia. Project included scraping the data from the various social media sites, building the data storage and deploying Power BI dashboard.
See Dashboard
The aim of this data science & machine learning project is mainly to predict the performance and math grades of students based on several quantitative and also socio-economic attributes. The project itself consists of data exploration, data analysis, data cleaning, transformations, feature engineering, construction of machine learning models and performing successful prediction.
See NotebookA fun web based game where you play as Baby Yoda. You collect stars and other items while avoiding bombs rushing all around the field. Built using HTML, CSS, and JavaScript.
Over 7+ years in the field, from traditional ML to modern GenAI / LLM applications. I design, build and bring AI solutions into production to help businesses increase profitability and improve customer experience.
LinkedinDuring the work on many projects, I have worked with a wide range of technologies and tools. I am always looking for new opportunities to learn and improve my skills.
GitHub