Skills
A brief summary of my skills and experience.
Applied statistics/Data science
- Time-series modelling (statistical and machine learning models).
- Bayesian and frequentist statistics (regression, hierarchical models, hypothesis testing)
- Extreme value analysis (from 1 to 10000 years return period events, stationary and non-stationary processes, GEV, GP, etc. with special emphasis on fitting models to relatively small datasets using Bayesian statistics).
- Deep learning: regression and image analysis (CNN).
- Risk assessment and decision theory.
Computer skills
- Python: numpy, numba, xarray, scipy, sklearn, pandas, matplotlib, seaborn, pyMC3, nestle, emcee, dynesty, pyABC, tensorflow, keras, pytorch, GPyTorch, ax, jMetalPy, OpenTURNS, catboost, lightgbm, sphinx, requests, flask, jinja2, etc., functional and object oriented programming, and code quality and package development tools.
- Julia: Flux.jl, DifferentialEquations.jl, DiffEqFlux.jl, Turing.jl, NLOpt.jl, GalacticOptim.jl, Zygote.jl, Test.jl, Pluto.jl, etc., and code quality and package development tools.
- R: dplyr, rstan, bmle, evd, parallel, shiny, ggplot2, leafletr, ggmap, renv, tibble, readr, etc.
- MATLAB: Parallel Computing Toolbox, Statistics Toolbox, Neural Network Toolbox, Optimization Toolbox, FERUM, UQLab, Advanpix, etc.
- AWS (EC2, S3), Azure (VM, DevOps, Logic Apps, Storage), Heroku.
- Git (GitLab, GitHub (both with CI/CD)), Docker, Jenkins, Terraform.
- SQL, PostgreSQL.
- GIS (QGIS).
- css, html, jekyll.
- ANSYS, MIDAS Civil, OpenSees, Diana, Atena, AxisVM, ConSteel.
- Linux, LaTeX, Microsoft Office products, AutoCAD, Adobe Illustrator.
Numerical analysis/Computational science
- Optimization: gradient-based (using finite differencing and algorithmic differentiation) and gradient-free methods as well, e.g. Newton-Raphson, Nelder-Mead, BFGS, COBYLA, stochastic gradient descent, Bayesian optimization (with surrogates), heuristic methods, multi-objective optimization, etc.
- Finite element method for solving PDEs.
- Computational methods for Bayesian parameter estimation and model selection (asymptotically exact and approximate methods for computationally demanding likelihood functions, MH-MCMC, variational Bayes (ADVI), nested sampling, NUTS, affine invariant ensemble sampling).
- Scientific machine learning (combining scientific models (e.g. in the form ODEs and PDEs) with machine learning models (e.g. neural networks)).
- Surrogate modelling (meta modelling) of computationally demanding models (e.g. computational physics simulations).
Leadership
- Technical lead and research proposal writing for >1 million € budget projects.
- Leading and managing multiple research teams of about 3-5 people.
- Planning and providing internal educational activities to colleagues on: software development, programming, applied statistics, risk and reliability analysis.
- Mentoring and supervising MSc students (>30).