Book · published by Sofixit
Big Book of Real-World Use Cases in Data & AI
Over 400 pages on mistakes, trade-offs and architecture from real deployments, written by 31 practitioners. In Polish, with no theory and no marketing gloss. The first print run has sold out - leave your address and we will write when the reprint or the e-book is ready.
authors and chapters
pages
parts: stories, case studies, interviews
published 18 March, in Polish
Who it is for
People who have already shipped something and want to compare their decisions with other practitioners.
Junior / Mid
Skip three years of mistakes in eight evenings of reading.
See production systems from the inside before you have to fix one yourself.
Senior / Architect
See the trade-offs in systems at scale.
Compare your decisions with those of 31 industry leaders and meet architectures that tutorials never show.
Manager / CTO
Learn how to shape Data teams so they deliver a return.
Not running projects but delivering value. Includes the plan "The first 100 days of a Head of Data".
This is not the book for you if…
- you are looking for Big Data and AI basics or an introduction for beginners. It is written in the specialised language that data and AI professionals use every day.
What is inside
Three parts: expert stories, real case studies and in-depth interviews. Every chapter stands on its own - read in any order.
01
Part I: Expert stories
Personal accounts of selected topics in a practical light: real architectural decisions, their consequences and lessons from the life of data projects.
02
Part II: Case studies
Real deployments, including BCNN (formerly Bielik Consulting) and Cyfronet: concrete business problems, the solutions applied, results and lessons.
03
Part III: Expert interviews
Conversations with practitioners about AI in everyday operations, data platforms, DataOps, Data Lake versus Lakehouse and the myths of data management.
- 01
Chapter 1: The first 100 days of a Head of Data
A ready plan: how to arrange the team, the technology and the relationship with the business to deliver results.
- 02
Chapter 7: Calls under the AI lens
100% of calls assessed in minutes by combining AWS Transcribe with language models.
- 03
Chapter 8: 69% bots in the data - how to fix it?
The fight for survey data quality: detecting low-effort answers and bots, a consistent category taxonomy, versioned models and prompts.
- 04
Chapter 12: Handling 8 billion events a day
A case study on scaling and cost optimisation. Concrete numbers and architecture.
- 05
Chapter 18: Knowledge graphs in investigations and medicine
How PDFs, tweets, logs and images become one coherent knowledge graph. A tool for investigative and medical analysis without the cloud.
- 06
Chapter 20: Not every problem needs an LLM
Matching the tool to the task: classification, anomaly detection, computer vision, route optimisation - and the usual traps of "LLM first".
Questions it answers
How to scale without going broke?
The architecture of a system handling 8 billion events a day and where to look for savings in the cloud.
Will your GenAI get past the tests?
Real deployments of language models, including Bielik, that left the trial phase and started earning.
Data Lake, Mesh or Lakehouse?
Why leaders chose specific solutions and which trade-offs they had to accept - instead of chasing buzzwords.
How to rescue data quality?
Why ML models break after a week on real data. A case study on filtering bots across 12,000 stores.
The authors on the book
Short recordings in Polish: what the authors contributed and why the book is worth your time.
Tomasz Tarczyński
Head of Data Science, IPF
Krzysztof Jędrzejewski
AI R&D Director, Pearson
Marcin Grzelak
Data Analytics Manager, Travelist
Patryk Hyla
Head of IT & Data, Danone
Kinga Barczewska-Pflanz
Head of Tech & Data, In-Pulse
Michał Krzyżanowski
Head of Data & AI, KUBO
Michał Domański
Co-founder, BCNN (formerly Bielik Consulting)
Filip Dzięcioł
Senior Data Engineer, Billennium
31 experts share what they know
Practitioners, leaders and solution architects who work with data every day in organisations around the world. Each brings a different career path, organisational culture and project scale.
Patryk Hyla
Head of IT & Data
Danone
Krzysztof Jędrzejewski
AI R&D Director
Pearson
Renata Bratkowska
Chief Data Officer & AI
NEUCA Group
Mateusz Jakitowicz
Chief Data Officer
Autopay
Kinga Barczewska-Pflanz
Head of Tech & Data
In-Pulse
Michał Szczudlak
Head of Data Analytics
G2A
Wojciech Piątkowski
Director of Data Analytics
Spyrosoft
Jakub Czerwonka
Chief of Data
Autopay
Michał Krzyżanowski
Head of Data and AI
KUBO
Marcin Grzelak
Data Analytics Manager
Travelist
Tomasz Tarczyński
Head of Data Science
International Personal Finance Plc
Paweł Gołąb
Co-founder
BCNN (formerly Bielik Consulting)
Michał Domański
Co-founder
BCNN (formerly Bielik Consulting)
Paweł Zieliński
CFO / CDO
Marek Stępień
Senior Manager
Mateusz Paździor
Head of Data Engineering
The Scalable Way
Emil Podwysocki
Deputy Head
National Information Processing Institute (OPI)
Maria Bylina
Data Governance Lead
National Information Processing Institute (OPI)
Kamil Raróg
CPTO
TelcoBrain
Błażej Ksycki
Head of Data
Roksana Wilk
Head of Data
ACK Cyfronet AGH
Małgorzata Belka
Senior Product Manager
Capgemini
Oliwia Walewska
AI Consultant
Rafał Wojdan
Data & AI Advisor
Onex Group
Aleksander Zawalich
Solutions Architect
Krzysztof Wasielewski
Chief Operating Officer
Sofixit
Filip Dzięcioł
Senior Data Engineer
Billennium
Marta Grzybowska
Team Leader
InPost Group
Jolanta Szymaniuk
Analytics Leader
Booksy
Kacper Grzyb
Data Architect
Sofixit
Michał Choiński
IT leader, consultant
The authors work at
- …and 12 more organisations
Reviews
Pull quotes translated from Polish; the full reviews are on the Polish page.
Marek Wilczewski
Managing Director for Information, Data and Analytics, PZU Group
„I feared that over 400 pages of specialist knowledge would chain me to the desk for at least eight long evenings. To my surprise, two were enough.”
Arkadiusz Dąbkowski
Enterprise Data & Data Governance Architect, The Stepstone Group
„This is a book about it not being easy. It tells what remains when a demo has to become a system: processes, data, responsibility and people's work.”
Weronika Dardzińska, PhD DVM
Real World Evidence Manager & Medical Affairs, AstraZeneca
„Written by practitioners for practitioners - people who work with data, build Data & AI solutions and face project constraints.”
Beata Buryn
Director of the Data Management Office, PEKAO
„The authors focus on concrete deployments, showing how Data & AI initiatives work in a real organisational setting, not only in theoretical models.”
Excerpts
Three chapter openings, as they appear in the book (in Polish).
Part II: Case studies
From an open question to an insight: AI in In-Pulse's qualitative analysis
In-Pulse runs shopper-led marketing research on data from more than 12,000 Żabka stores, about 4.2 million transactions a day. Coding open-ended survey answers used to take the team up to two days per project; with GenAI it takes minutes, with people staying in the loop for validation and interpretation.
The full 11-page description of the deployment is in the book.
Part I: Expert stories
The first 100 days of a Head of Data: three axes that define success
The first 100 days are the moment of truth for every Data & Analytics leader: credibility with the board, the team's role in the organisation, and whether Data will be a support function or a strategic partner. Success depends on balancing three axes - operational, strategic and team. Day one rule: protect what already works.
The chapter continues with the strategic and team axes and a 100-day plan.
Part I: Expert stories
Compose Your Next Reach: how Autopay changed ad targeting with AI
Autopay processes over 160 million transactions a year and holds close to 300 data points for each of more than 6 million users - yet the sheer volume had become a barrier for marketers. The answer was Composer, a data-driven tool that uses AI.
The chapter continues with how Composer was built and what it changed in campaigns.
Why does a technology company publish a book?
At Sofixit we build data and AI solutions for large organisations in Poland and Europe. Working with clients we see one thing: experts need practical knowledge, not another theoretical course.
When we started talking to the best specialists in the field, they all shared the same observation - there is no place where real case studies are shared. Not the successes from presentations, but the real decisions, mistakes and lessons.
„We are not a publisher, but we know what working with data looks like in practice. So we decided to create what we were looking for ourselves - a compendium of knowledge from people who really solve problems in their organisations.”
Print run sold out
Reprint and e-book: the waitlist
The first print run has sold out. Leave your address and we will write when the reprint starts or the electronic edition appears. Nothing else - no newsletter.
Bulk orders for teams and companies at the reprint: [email protected].
Questions buyers ask
Can the book still be bought?
The first print run has sold out. Join the waitlist - we will announce the reprint and the e-book as soon as they are available.
Will there be an e-book?
We are working on it, but there is no date yet. People on the waitlist will hear first.
What language is the book in?
Entirely in Polish.
What exactly does it contain?
Over 400 pages, 31 chapters in three parts: expert stories, case studies of real deployments and interviews with experts. Every chapter stands on its own.
Can we order copies for a team or a company at the reprint?
Yes. Write to [email protected] - we will agree the number of copies and the VAT invoice.
Where can I watch the BIG BOOK LIVE conference?
The recording of the conference of 25 September 2026 is free. Register briefly on the BIG BOOK LIVE page and we will e-mail you the link. The agenda and the speakers are there too.
Who published the book?
Sofixit sp. z o.o., Wrocław. Published on 18 March 2026.
BIG BOOK LIVE · 25 September 2026
The conference that grew out of the book
Four panel discussions by the authors and reviewers: data and AI from practitioners for practitioners, European infrastructure, technical debt and lessons from AI deployments. The recording runs over four hours and is free after a short registration (in Polish).
Watch the recording