Machine Learning / Data Science
Andrey
Musatov
Data Science MSc candidate and Computer Science graduate (2:1) with experience supporting undergraduate machine-learning teaching. I build practical ML and data-driven systems, contribute to open source, and care about rigorous evaluation as much as implementation.
PyTorch
scikit-learn
pandas / NumPy
Git / Linux / Docker
Education & languages
Education
MSc Data Science
Lancaster University Leipzig · In progress
Developing deeper expertise in machine learning, statistical analysis, data visualisation, model evaluation and applied data-science workflows.
BSc Computer Science
Lancaster University Leipzig · Graduated 2026 · 2:1
Relevant study included statistical research methods, evaluation of AI systems and modern communication protocols, alongside machine learning and software engineering.
Languages
Technical skills
Machine learning & data
- Classification & regression
- Feature engineering & preprocessing
- Model evaluation & statistical analysis
- Recommender systems
- Computer vision
Scientific Python
- Python
- PyTorch
- scikit-learn
- pandas & NumPy
- OpenCV
Engineering
- Git & GitHub workflows
- Linux
- Docker
- Django
- SQL
Current focus
- Applied machine learning
- Time-series & probabilistic modelling
- RAG & embeddings
- LLM evaluation
- Reliable model behaviour
Experience
Lancaster University Leipzig
Teaching / Research Assistant
October 2025 — May 2026
Supported machine-learning classes for approximately 20 undergraduate students, covering supervised and unsupervised learning, data manipulation and core ML concepts. Provided academic support and helped facilitate discussion around technical course material.
machine learning
data manipulation
academic support
Selected work
Projects selected for the problems they tackle and the technical decisions behind them.
Recommender Systems / Machine Learning
AniRecom ↗
Adaptive recommendation-system experiment that turns anime metadata into machine-readable features and uses explicit user feedback to influence subsequent recommendations.
categorical features
user feedback
data processing
Django / Security / Traceability
MintDPP ↗
Digital Product Passport MVP exploring product transparency, traceability and authenticity, combining a Django application with cryptographic mechanisms for trustworthy product records.
cryptography
data integrity
web application
Machine Learning / Risk Analytics
SmokyAlert ↗
Experimental company-risk modelling using labelled numerical data, temporal event counts, sentiment indicators and operational signals. The workflow focuses on feature engineering, leakage-aware evaluation and interpretable model performance.
scikit-learn
feature engineering
model evaluation
Open source
Upstream contributions show how I work in unfamiliar codebases, respond to review and ship changes others rely on.
Merged · PyMC Extras · PR #757
Multivariate log-likelihood fix ↗
Corrected the multivariate SquareRootFilter log-likelihood calculation and added regression coverage against the expected Gaussian result.
numerical correctness
regression testing
maintainer review
Open PR · aeon · PR #3752
dtype preservation in time-series transforms ↗
Working on dtype preservation across multiple time-series transformers to avoid unnecessary float32-to-float64 promotion, with regression tests and active maintainer review.
time-series ML
dtype handling
test design
Merged · Clips Studio · PR #66
Russian localisation completion ↗
Completed and corrected Russian localisation for an open-source local-first AI video application; the contribution was the project's first external pull request.
QA / consistency
upstream workflow
Merged · YazSes · PR #260
First Cyrillic README translation ↗
Added Russian documentation and language navigation to a privacy-first offline voice-dictation project, contributing its first Cyrillic-script translation.
localisation
contributor workflow
Contact
Open to graduate and early-career opportunities in data science, machine learning and applied AI.
Especially interested in roles where careful experimentation, model evaluation and software engineering come together to solve practical problems.