Daniel Wlazło¶

Data Science Manager at Allegro Pay — trustworthy machine learning for credit risk, in Warsaw.

from datadeer import Daniel

d = Daniel(role="Data Science Manager", org="Allegro Pay", city="Warsaw")
d.focus     # ["application scoring", "fairness", "calibration", "explanation"]
d.thesis    # "discriminatory uncertainty" (SGH Warsaw, in progress)

About¶

I’m a Manager of Data Science in credit risk at Allegro Pay, where I lead modelling work for one of Central Europe’s largest consumer‑finance platforms. My focus is application scoring — PD models for individual borrowers — and, increasingly, what trustworthy machine learning looks like in a regulated domain: fairness, calibration, explanation.

I’ve spent close to a decade in credit risk and data science, across Allegro Pay, Hexaware (for AXA), PKO Bank Polski, and ING Bank Śląski. Before that I was a robotics engineer at Samsung and a UAV constructor at Aviation Technik. Robotics was where I started — it’s still how I think about machine learning: a system you can describe, debug, and put under load.

Outside work I learn Czech, follow Formula 1, and tinker with robotics.

Didactics¶

Teaching materials: github.com/wlazlod/didactics

Conformal Prediction in Credit Risk — Research Workshop¶

Warsaw University of Technology with Allegro Pay

Industry‑mentored project for the Research Workshop course at the Faculty of Mathematics and Information Science. Three companies contribute a project each year and guide the student team through it; ours is on conformal prediction applied to consumer‑credit scoring. I co‑prepared the brief and mentor the team.

Starter repository · Teaching materials

Academic work¶

Doctoral thesis — Trustworthy credit risk under the AI Act¶

SGH Warsaw School of Economics in progress submission target 2029

Supervised by Prof. Aneta Ptak‑Chmielewska. The thesis develops an integrated methodology that combines algorithmic fairness, uncertainty quantification, and interpretability, organised around a concept I call discriminatory uncertainty — the idea that a model can discriminate not only through its decisions, but through how confident it is in them.

Open source¶

treecf¶

python rust core pre-release v0.1 MIT

Constrained, threshold‑aware counterfactual explanations for tree ensembles (XGBoost, LightGBM, CatBoost, sklearn): the minimal, feasible change that moves a declined application under the cutoff. Plans respect declarative real‑world constraints, treat missing values as first‑class, and every answer is float‑verified against the model before it is returned. The search runs on a Rust engine bundled in the wheel — milliseconds per applicant, whole portfolios in one parallel call.

Documentation · GitHub

concept-graph-xai¶

python alpha v0.2 MIT

A library that maps a tabular model’s raw features into business‑level concepts and visualises how much of each concept the model actually uses. Includes concept‑level ablation (permutation, SHAP‑marginal, retrain), missingness diagnostics, and regulatory tag overlays.

Documentation · GitHub

swift¶

python MIT

SHAP‑Weighted Impact Feature Testing for ML system monitoring. Weights feature drift by SHAP‑derived impact, so the alerts you get are the alerts that actually move the model’s output.

Documentation · GitHub

triadxai¶

python v0.1 Apache-2.0

TRIAD — a three‑channel decomposition of model predictions for credit‑risk explainability. Each feature’s contribution is split into Information, Density (epistemic), and Missingness channels that sum back exactly to the score, separating “denied because of bad information” from “denied because we lack information”. Exact for additive models, SHAP‑based for black‑box GBMs.

Documentation · GitHub

flaggam¶

python Apache-2.0

Interpretable generalized additive models built from univariate flags. A screening pass finds the threshold and category cuts where a feature’s distribution shifts against the outcome; each cut becomes a binary rule, and a linear head fitted on top gives exact rule extraction and feature‑level attribution. Sklearn‑compatible estimators.

Documentation · GitHub

Conference talks¶

Slides and materials: github.com/wlazlod/conference-talks

• The Scorecard Illusion: Why SHAP values are not points (and how they can fool you) — Data Science Summit, AI Edition 2026, Warsaw.
  Why a SHAP value behaves like a contribution, not a coordinate — and how reading it as a point misleads scorecard interpretation. Slides (PDF) · write-up.
• Curing Semantic Debt: Practical Applications of Knowledge Graphs in ML — GHOST Day 2026, Poznań, with Kamil Golis.
  How knowledge graphs pay down the semantic debt that accumulates across ML pipelines — a practical, applied walk‑through.
• Poza Feature Store: jak Ontologie Semantyczne mogą przyspieszyć rozwój modeli ML — Data Science Summit 2025, Warsaw.
  How semantic ontologies can move ML development past the feature store, giving models a shared, reusable vocabulary.

Notes¶

• 2026 Discriminatory uncertainty
  A model can discriminate not only through its decisions, but through how confident it is in them.
• 2026 Why SHAP values are not points
  A SHAP waterfall looks like a scorecard. Read it like one and it hands out advice that is confidently wrong.

Elsewhere¶