Oxford Mathematical Physics DPhil candidate with experience in Python, statistical modelling, machine learning and optimisation. I build structured, reproducible modelling workflows and decision-support tools, with projects spanning ML pipelines, constrained optimisation, time-series backtesting and interactive simulation.

I am interested in roles across data science, applied analytics, quantitative modelling, optimisation, and research/analytics engineering.

Python decision-support pipeline for the official F1 Fantasy game. The project ingests historical race data, builds recency-weighted expected scoring estimates, and solves team selection under budget, roster and transfer constraints using mixed-integer optimisation.

Relevant skills: Python, pandas, PuLP, constrained optimisation, feature construction, decision modelling, explainable recommendations

Repo: f1_fantasy_optimizer

End-to-end machine learning pipeline for credit-risk classification, from data preparation to model comparison and threshold tuning. Models include logistic regression, random forest and XGBoost, evaluated using cross-validation and PR-AUC to handle class imbalance.

Example results: ROC-AUC ~0.77, PR-AUC ~0.55; minority-class recall improved from ~35% to ~59% through imbalance-aware evaluation and threshold optimisation.

Relevant skills: Python, pandas, scikit-learn, XGBoost, cross-validation, model evaluation, imbalanced classification, threshold optimisation

Repo: ml_pipeline

Reproducible research and backtesting framework for testing conditional weekly market patterns. The project separates exploratory analysis from strategy evaluation and includes risk-based sizing, layered exits and drawdown diagnostics.

Relevant skills: Python, pandas, time-series analysis, backtesting, performance evaluation, risk diagnostics, reproducible research structure

Repo: monday_range

Lightweight browser-based simulation of radar sweep detection, target acquisition and dynamic pursuit assignment. Built as an interactive demo to prototype decision logic for multi-target tracking and interception scenarios.

Relevant skills: JavaScript, HTML Canvas, simulation logic, real-time visualisation, target assignment

Repo: drones

Programming: Python, Git, Jupyter; currently learning SQL; working knowledge of JavaScript/HTML, MATLAB and C++

Python stack: pandas, NumPy, SciPy, scikit-learn, XGBoost, PuLP, Matplotlib

Modelling & Analysis: statistical modelling, regression, classification, cross-validation, model evaluation, threshold optimisation, feature engineering, imbalanced classification, time-series backtesting, constrained optimisation

Mathematics: linear algebra, probability & statistics, regression modelling, numerical methods, optimisation, stochastical processes

• GitHub: https://github.com/dnpjr
• LinkedIn: https://www.linkedin.com/in/daniel-pajer/
• Email: dancer-bout.6v@icloud.com

This profile is a living portfolio: projects and write-ups are being expanded with robustness checks (costs/slippage, walk-forward validation, sensitivity analysis) and additional modelling examples.