Actuarial Analyst | Econometrician | Quantitative Risk Modeller

I work at the intersection of actuarial modelling, econometrics, financial risk and applied data analysis. My current professional focus is life-insurance model testing and validation, with prior actuarial consulting experience in pension and employee-benefit valuations under HGB/IFRS.

My GitHub portfolio is built around quantitative projects in insurance analytics, financial econometrics, survival analysis, Bayesian modelling, portfolio optimisation and machine learning. The repositories are primarily academic or analytical projects, with emphasis on reproducible workflows, transparent modelling assumptions and careful validation.

• Actuarial modelling: life insurance, cash-flow projections, Prophet workflows, pension valuations, KNF actuarial exam progress.
• Financial and insurance risk: GARCH and Markov-switching GARCH models, VaR, Expected Shortfall, backtesting, telematics insurance.
• Statistical modelling: survival analysis, Bayesian regression, credit scoring, classification, model diagnostics.
• Optimisation and simulation: Markowitz portfolio optimisation, cardinality constraints, metaheuristics, Monte Carlo simulation.
• Data engineering for analytics: Kafka, Spark Structured Streaming, FastAPI, reproducible Python/R notebooks.

Course-oriented usage-based insurance project using synthetic telematics events, Kafka, Spark Structured Streaming and an interpretable Poisson GLM with exposure offset. The project separates periodic technical premium scoring from real-time operational decisions such as safety nudges and driver-status checks.

Academic optimisation project for the Markowitz portfolio problem with a cardinality constraint. It compares Simulated Annealing, Genetic Algorithm and Particle Swarm Optimisation, with portfolio risk evaluated through Monte Carlo scenarios generated from a GARCH(1,1)-based process.

Survival analysis project using clinical heart-failure records with right-censored observations. It combines Kaplan-Meier and Nelson-Aalen estimators, AFT models, a demonstrative Bayesian Weibull model, Cox proportional hazards modelling and Schoenfeld-residual diagnostics.

Bayesian modelling workspace based on the Adult Census Income dataset. It compares a Bayesian linear probability model with Bayesian logistic regression for modelling the probability of income above USD 50,000.

Data mining project for binary passenger-satisfaction classification. It compares logistic regression, random forest, histogram gradient boosting and multilayer perceptron models, with scikit-learn pipelines, model diagnostics, permutation importance and error analysis.

Languages and analytics: Python, R, SQL, MATLAB, SAS, VBA, Markdown, LaTeX, Typst
Python stack: pandas, NumPy, scikit-learn, statsmodels, lifelines, PySpark, FastAPI
Actuarial and finance: Prophet, Retirement Studio, Excel/VBA, HGB/IFRS Methods: GLM, logistic and linear regression, classical machine learning, explainable AI, Bayesian modelling, survival analysis, credit scoring, optimisation, Monte Carlo simulation

• Prophet-based actuarial model testing and validation in health insurance.
• MSc studies in Advanced Analytics - Big Data and Machine Learning at SGH Warsaw School of Economics.
• Organising a public portfolio around actuarial analytics, econometrics and quantitative risk modelling.

• Website: jakubmikolajczak.pl
• LinkedIn: Jakub Mikolajczak
• GitHub: jmDarcy
• Email: jakob.mikolajczak@gmail.com