AI & LLM Consultant | Data Scientist | Mentor
Meet your go-to AI & LLM Consultant and Data Scientist with over a decade in the analytics trenches.
Throughout my career, I've led cross-functional teams to deliver AI-driven projects in banking, insurance, and beyond. My solutions have trimmed response times by half, boosted fraud detection accuracy beyond 85%, and generated 200+ successful customer campaigns.
And yes, I keep things fun along the way. Whether it's integrating large language models into your workflow or demystifying analytics for business stakeholders, I'm all about practical, results-oriented data science - sprinkled with a dash of humor. Let's analyze, optimize, and watch your business prosper.
Contact me at: [email protected]
With a passion for analyzing data and turning it into actionable insights, I thrive on delivering end-to-end machine learning applications. My mission? Bring cutting-edge AI solutions to life and help businesses excel in today's data-driven landscape.
Currently, I'm diving deep into large language models (LLMs) to build next-generation prototypes with GPT & Agents. I've developed real-time analytics pipelines, implemented advanced risk scoring models and led global teams in fully remote environments.
Feel free to connect with me on LinkedIn, check out my code on GitHub, or explore the latest insights on my blog.
Senior Analytical Consultant
SCS Expert Sp. z o.o., Warsaw, Poland
01.2022 - Present
Analytical Consultant
SAS Poland Sp. z o.o., Warsaw, Poland
01.2020 - 12.2021
Data Analytics Capability Manager (Assistant Vice-President)
Citi UK, London, United Kingdom
07.2018 - 06.2019
Data Analytics Head
Citi Indonesia, Jakarta, Indonesia
01.2018 - 06.2018
CRM & Campaigns Head
Citi UK, London, United Kingdom
09.2015 - 06.2018
Customer Analytics & CRM Expert Analyst
Citi Poland, Warsaw, Poland
06.2014 - 08.2015
CRM Specialist
Compensa Vienna Insurance Group, Warsaw, Poland
10.2012 - 05.2014
Economics of Entrepreneurship
University of Warsaw (21/22, 2024 - 2025)
Data Science in Business
University of Warsaw (2019 - 2020)
Club Member
Toastmasters Leaders (2013 - 2014)
Computer Science & Econometrics
University of Warsaw (2008 - 2014)
Realized projects, proof-of-concepts and prototypes.
Struggling to price properties accurately or losing revenue due to poor forecasts? This project demonstrates how a robust Theil-Sen regression pipeline can enhance property valuation by handling outliers effectively and offering more reliable predictions of house sale prices. By leveraging advanced feature engineering (like Weight-of-Evidence and Box-Cox transformations), this solution helps real estate professionals, investors, and analysts make data-driven decisions that optimize pricing strategies, boost market insights, and reduce financial risk.
I led the end-to-end development of this model—covering everything from data cleaning and transformation to hyperparameter tuning—so you can confidently deploy a battle-tested regression solution. This includes detailed exploration of data, systematic feature selection, and a robust model evaluation framework that ensures high accuracy and actionable predictions for your real estate portfolio.
Do you find your data science projects turning into a chaotic tangle of scripts and folders? Smart Structuring offers a step-by-step guide to organizing code, managing dependencies, and maintaining version control. Whether you're a startup striving for reproducibility or a large enterprise looking to streamline collaboration, this e-book provides real-world best practices and proven tips to minimize errors, speed up development cycles, and ensure scalability.
Drawing on my experience as both a data scientist and technical writer, I compiled actionable insights to help you avoid common workflow pitfalls - from inconsistent naming conventions to untracked data changes. The e-book is engineered to give you a reliable roadmap for building robust, maintainable projects that withstand the complexities of modern data science and keep your team focused on results instead of overhead.
Tired of bland, scripted chatbots that barely engage your users? This playful AI comedian breathes new life into customer support, marketing campaigns, or product demos by delivering witty jokes and spontaneous quips. Leveraging advanced NLP and deep learning—specifically the LLama 3.1 Instruct model with 8 billion parameters—this project demonstrates how entertaining interactions can boost user satisfaction and increase brand engagement.
I handled the full-stack deployment, blending large language model technology with an intuitive front-end to ensure real-time performance and an engaging user experience. By bringing humor into the conversation, the bot enhances customer engagement and can even serve as a creative marketing tool, illustrating how cutting-edge AI goes beyond pure utility to foster memorable brand experiences and loyal customer relationships.
Want to see what you can achieve with AI? Get in touch! Whether you're taking your first steps in data science, creating prototypes, optimizing ML pipelines, or want to reach for innovations with LLMs — just one conversation stands between you and getting significantly closer to achieving your goals!
Explore a variety of AI and data science articles, each delving into practical tips, real-world scenarios, and best practices. Stay informed, stay empowered!
A hands-on introduction to Claude Code as an AI pair programmer, with setup tips, workflow patterns, and hard-won lessons from a two-week refactor sprint. Includes guidance on planning, testing, and keeping control over agentic tools.
A practical introduction to getting better results from LLMs: free ways to compare models, meta-prompting, a simple prompting framework, and Socratic questioning. Part of the Guardrail Garage series.
A practical tour of tokenisation for LLMs. Why word-level fails, how subwords and BPE work, and what modern tokenisers look like, with tips and resources. Plus a brief personal update.
An accessible tour of word embeddings, from Word2Vec to GloVe, with intuition, a touch of maths, and practical notes on training and using them. Includes caveats, examples, and where modern models fit.
