Piotr Żak
Lead Data Scientist
An Oxford-trained physicist and generalist data scientist: across 13 years I have owned the full data science process – data discovery, architecture, modelling, deployment and handover – on problems ranging from forecasting and predictive maintenance to NLP and propensity modelling. That breadth stays current with the field, agentic GenAI included: multi-agent systems, RAG pipelines and MCP tooling delivered as production-grade products. I lead both the teams that build these systems and the stakeholder conversations that shape them.
Selected Professional Experience
SoftServe
Remote, Poland
Lead Data Scientist
Jun 2024 – Present
- Led development of a GenAI document-intelligence platform for a global pharmaceutical manufacturer – a modular LLM pipeline that extracts and classifies requirements from technical PDFs, resolves them against engineering diagrams using a vision-built connectivity graph, and formalises them into machine-executable rules (Azure OpenAI, Azure Document Intelligence, Streamlit, Azure DevOps).
- Led delivery of a GenAI negotiation-support platform for a TOP3 global management consultancy – defined the scope and delivery plan directly with the client and led a team of 15 engineers across UI, GenAI, data-pipeline and QA streams. Built through an LLM-based delivery system designed to keep the human in the loop to a minimum, shipping 36 screens and 4 GenAI use cases in 15 weeks (Figma-to-code generation).
- Built a GenAI application that generates PRDs, epics, user stories and tasks from user input, using specialised agents integrated with MCP tools. Implemented the RAG system, agentic workflow, MCP servers and frontend (OpenAI Agents SDK, FastAPI, Pydantic, Redis).
- Developed models estimating the probability of oil-rig component failure from unstructured text in failure reports, for a Norwegian offshore oil and gas operator. Led data discovery, scoped the long-term modelling plan and built a neural network with an NLP component (Hugging Face, Databricks, PySpark, PyTorch, Power BI).
Senior Data Scientist
Aug 2023 – Jun 2024
- Served as Lead Data Scientist on a programme spanning 50+ technical professionals, delivering predictive maintenance for one of the largest windfarms in Europe. Deployed two anomaly-detection models to production, helped stand up the MLOps platform and drove business and data discovery (Databricks, PySpark, Azure ML, Power BI).
- Led an international team of 5 to deliver slime-level forecasting for a world-leading producer of mineral sands in Mozambique. Shipped a forecasting module built on NeuralProphet with week-to-week what-if analysis, a Power BI dashboard, a dedicated database and a simplified MLOps setup; owned scoping, architecture, modelling, client workshops and the full codebase.
Algomine
Remote, Poland
Data Scientist
Sep 2021 – Aug 2023
- Led a project for a major Polish media agency using off-the-shelf computer-vision models (Deepface) to estimate the age of people in images scraped from multiple search engines, quantifying the gap between media-portrayed and real age across professions and geographies (GCP).
- Contributed to the client's open-source ML algorithm inferring race/ethnicity from only a full name and US home address — used to detect racial bias in credit-scoring models. Refactored and sped up key components and trained models on ~2 TB of data (machine with ~4 TB RAM).
- Built a propensity-to-buy model for credit products of the corporate division of a large Polish bank, owning the project from sales process through to product handover.
- Built a demand forecasting system covering 200 SKUs across 30 distribution channels for a leading Polish pharmaceutical manufacturer; joined as a team member and was quickly moved into the lead role. Owned data collection and staging, feature engineering, model training (gradient-boosted trees), visualisation and client handover (PySpark, Azure, Databricks, CI/CD).
Warsaw, Poland / Remote
Polish (native), English (fluent)
Technologies
Machine learning
ML algorithms (XGBoost, LightGBM, NeuralProphet, neural nets and more)
scikit-learn
PyTorch
Optuna
SHAP
A/B testing
MLOps
Model monitoring / drift
Advanced maths
GenAI / NLP
OpenAI Agents SDK
LangChain / LangGraph
RAG / MCP
Skills / scripts
Hugging Face Transformers
Embeddings
Vector databases (e.g. Redis)
Prompting / guardrails
FastAPI / Pydantic
Engineering
Python (incl. OOP)
SQL
PySpark
Databricks
Docker
Cloud
Git
CI/CD
MLflow / Azure ML
Azure ML Pipelines / Databricks Jobs
Power BI
Santander Bank Polska
Warsaw, Poland
Product Owner, Analytical CRM / Data Scientist
Jun 2020 – Jul 2021
- Designed the infrastructure for the bank's proprietary analytical CRM system and led a team of developers and data engineers (10–20) responsible for the implementation. Rolled out 30+ sales campaigns, ranging from rule-based targeting on selected features to dedicated propensity-to-buy models.
Senior Data Analyst / Data Scientist
Nov 2017 – Jun 2020
- Responsible for the department's key analytical initiatives, creating business value from data through process optimisation and executive reporting.
The Boston Consulting Group
Warsaw, Poland / Glasgow, UK
Associate
Dec 2016 – Nov 2017
- Consulted a leading Polish fashion retailer on a data-monetisation strategy, covering data governance, demand forecasting and AI solutions to enhance sales and operations; led knowledge-sharing sessions with the client and delivered the final report.
PKP Intercity
Warsaw, Poland
Adviser to the CEO / Data Analyst
Aug 2014 – Apr 2016
- Advised the CEO on analytic and strategic decisions, including a PLN 1bn recapitalisation and a 30-person long-term strategy. Led analysis that cut standing passengers on Express InterCity trains by over 40% at no extra cost, and designed the 'Happy Hours' offer behind a 22% y/y rise in passengers.
Macquarie Group
London, UK
M&A Intern
Jun 2013 – Aug 2013
- Built financial models and pitch books, and wrote a client-facing report on the European telematics auto-insurance market.
Education
Warsaw University of Technology
Oct 2019 – Jun 2020
Postgraduate Studies – Data Science
University of Oxford
Oct 2011 – Jun 2014
BA Physics, St John's College – Final grade: 2:1
Community
Oxford Finance Society
Vice President
Led society programming and events across the University of Oxford.
ELAB Education Laboratory
Mentor
Supported numerous high-school students in securing places at the world's top universities.
Interests
TOP110 squash player in Poland
Passionate kite-surfer
Snowboarding instructor
Adventurous traveller