layout | title |
---|---|
cv |
Maciej Misiura CV |
[email protected] - +44 (0)7790 597 206
PhD-educated Data Scientist, specialising in statistics and machine learning, with proven analytical, communication and management skills developed in business and academic environments. In my current role, I help local, national and international companies to develop and implement data-driven solutions to their business problems.
2021-Now
Data Scientist, National Innovation Centre for Data, Newcastle upon Tyne
- Headed a project on training and validating a large language model with uncertainty quantification capabilities to improve performance and decision making for an intelligence and security organisation
- Managed a project on building interpretable tree-ensemble and uplift models to better understand customer churn and improve retention rates of a large software company's flagship product
- Led the design and development of a recommendation system to facilitate onboarding of new customers of a large software company
- Assisted clients in migrating their existing data science projects to Databricks
- Co-developed an end-to-end classification model ranking prospective customers of a leading real estate consultancy to improve the effectiveness of marketing campaigns
- Deployed machine learning models to production using different technologies, including Datarbricks and AWS Lambda
- Mentored team members on the use of Data Science Process frameworks, such as CRISP-DM and Data-Driven Scrum to efficiently manage data science projects
- Created and delivered teaching content to clients on:
- classification on imbalanced data
- interpretable machine learning
- causal inference in machine learning
- MLOps
- Produced several research reports for the Applied Research Centre for Defense and Security, including proof-of-concept work on:
- evaluating the performance of different uncertainty quantification methods with a transformer-based natural language model
- comparing the uncertainty-unaware few shot deep learning models to their uncertainty-aware, Bayesian counterparts
- identifying the state-of-the-art neural networks for key word identification in videos
2018-2019
Data Scientist Placement, Newcastle University, Newcastle upon Tyne
- Contracted with the the Agriculture and Horticulture Development Board to undertake an investigation into quantifying environmental impacts of the UK commercial livestock production systems over the last two decades
- Displayed strong analytical skills through the visualisation and interpretation of complex, commercial data
- Delivered a 10,000-word report, which should inform future government policy on identifying effective strategies to minimise environmental impacts of raising livestock
2015-2016
Indirect Tax Analyst, Deloitte, Newcastle upon Tyne
- Delivered tailored-made tax advice to clients from a variety of sectors including education, automotive, engineering and charity
- Improved technical writing skills by drafting engagement letters and client reports
- Demonstrated strong independent research skills and initiative by creating a large database of North East-based companies, which was utilised across regional offices for business development and client targeting
2016-2021
PhD Applied Mathematics and Statistics, Newcastle University, Newcastle upon Tyne,
Thesis: Mathematical and statistical modelling for a more efficient and sustainable pig production:
- Investigated ways of improving environmental sustainability of commercial pig production systems through an interdisciplinary approach encompassing computational biology, precision agriculture, nutrition, data science, systematic reviews and meta-analyses
- Presented research on smart agriculture at large international conferences, including talks at EAAP and ASAS
- Authored 6 scientific articles, including 4 first author papers
- Awarded the 2018 British Society of Animal Science Murray Black Award (£1,500) for an outstanding research proposal
2011-2015
Mathematics and Statistics MMathStat, 2:1, Newcastle University, Newcastle upon Tyne,
Thesis: Bayesian analysis of paired comparison data:
- Delivered a 5,000 word dissertation describing predictive models of basketball matches and entered the Kaggle competition to predict the outcomes of the March Madness tournament
Publication list is available on ResearchGate
NumPy, SciPy, pandas, scikit-learn, PyTorch, Hugging Face
Tidyverse, Tidymodels, jags, Stan
MLflow, Spark MLlib, XGBoost, Spark SQL
Git, GitHub