AI Validation Framework
Testable. Validatable. Trustworthy.
A pragmatic system that makes any AI implementation testable and validatable — raising data quality so that results can actually be trusted and relied on in production.
Software engineer and solution architect, 7 years across healthcare, data and gaming. At Holisticon I co-author the frameworks we use to roll AI into large, regulated companies — and I lead the work that gets them shipped.
I started as a backend developer and grew into the person who owns how AI actually lands inside a big company — the architecture, the data, the way a team works around it, and the trust that lets leadership say yes. I still write production Python, TypeScript and .NET; I just also shape the decisions around them.
At Holisticon I co-authored two frameworks that make AI adoption structured and testable, and I help large companies in regulated, data-critical fields put them to work. A Master's in Management on top of Computer Science is what lets me run a delivery and talk to a board in the same week.
Production Python, TypeScript and .NET across cloud, data and AI.
I design how AI systems are built, governed and trusted — then lead delivery.
IEEE-published; I turn messy AI problems into shareable structure.
I co-authored two frameworks that shape how AI gets built and adopted — and I put them to work with Holisticon inside large companies.
Testable. Validatable. Trustworthy.
A pragmatic system that makes any AI implementation testable and validatable — raising data quality so that results can actually be trusted and relied on in production.
AI as part of the team — not just a tool.
A model for how IT teams operate when AI is a first-class actor in the process — redefining roles, governance and the way work is delivered end to end.
[ i ] Both frameworks are developed at and owned by Holisticon. Shown here as authorship, not disclosure.
With Holisticon I implement AI systems inside large companies — across medical, data-analytics and beyond. The specifics stay confidential; the approach stays constant: governed, validated, trustworthy.
GenAI tooling and data-validation pipelines for a global medical company.
AI agents and validation layers for data-critical, high-stakes workflows.
Rolling out the AI-First way of working at organizational scale.
Beyond client work, I'm building and co-building a few things of my own.
A platform for searching and comparing real-estate information — pulling scattered property data into one place so you can compare and decide faster.
Want to test it? Let me know.A document-automation platform: build advanced data and document pipelines visually from modular AI blocks. Each block does one deterministic job — extract, split, analyze, report — and AI models slot in where they help, so processes stay composable and explainable. No-code, with a community block marketplace and on-premise scaling for enterprise.
doccompan.io ↗A research project turning WiFi sensing into a product: reading wireless signals to map spaces, build heatmaps and track moving objects. Dual-use by design — civilian and defense. Early stage, first proof of concept, built with people from Poland's Military University of Technology (WAT).
AI Implementation Lead · Solution Architect · Full-Stack Developer
I lead how AI systems are designed and rolled out inside large enterprises, and co-author two internal AI-delivery frameworks. I also work hands-on as a full-stack developer for a global medical client — building across Django + React and AI-agent tooling, focused on governed, validatable data pipelines for GenAI in pharma. Delivered a research paper and an industry talk on data validation for GenAI.
Full-Stack Developer · Gaming AI & Tools
Interactive web apps and AI-powered tools for mid-core RPGs: modular React + TypeScript UIs, .NET and Blazor services, real-time features and Python REST backends in Dockerized environments.
Python / Full-Stack Developer
Flask microservices with Vue.js frontends. Built ETL pipelines with PySpark, optimized system performance and delivered end-to-end test automation across cloud infrastructure.
Python Developer
Django, Flask and REST solutions with React and PostgreSQL. ElasticSearch, large-scale data operations, code reviews and Dockerized environments.
Web Engineer
Python REST APIs, system integrations and e-commerce website development with JavaScript.
Software Developer
Contract software for diverse clients — secure, scalable systems in Python and JavaScript, spanning web apps, automation and API integrations.
IEEE-published researcher in AI incident data sharing and governance. I take messy, real-world AI problems and turn them into structured, shareable knowledge — from conference papers to industry talks and technical articles.
Kozminski University
Business strategy, leadership and organizational management — the side that complements engineering when AI meets a real business.
University College Dublin
Data Science, web programming and Python, with research in AI, cybersecurity and analytics.
Certification in delivering AI projects with the CPMAI methodology.
pmi.org ↗ALX.pl
Microsoft Poland internship
Akademia 108
Marcel is a great software engineer and team member. He picks up new skills quickly and applies them. He can work individually or collaboratively — I can highly recommend him for any software engineering role. Excellent Python, CI/CD, Cloud and Vue skills.
Marcel is extremely good at dealing with ambiguity, quickly providing working code to discuss and iterate on. With stricter requirements he ensures the code is well tested. I highly recommend working with Marcel — problem solving, fast learning and clear communication with engineers and stakeholders.
Working with Marcel on "Recycle.it" and other projects was a fantastic experience. His expertise in Python and web development was key to our success, and his innovative solutions and leadership were invaluable. Highly recommended.
For AI implementation, architecture, or simply a conversation about doing AI properly — reach out.
@ kontakt@marcelkaminski.pl·⌖ Warsaw · Remote