I am a scientific Python developer and computational neuroscience researcher. I build tested data pipelines, ML evaluation workflows, and reproducible research software for complex time-series and neurodata.
My work combines software delivery with research practice: reproducing difficult bugs, turning analysis code into maintainable packages, implementing validation and quality-control workflows, and documenting the assumptions behind numerical results. I have a Master's degree in Psychology with Distinction from HSE University's Cognitive Sciences and Technologies programme, with training in EEG/ERP methods, neuroimaging, neural modelling, statistics, and experimental design.
I am available for remote contract and part-time work in scientific Python, ML evaluation, research software, and data pipelines. I am also open to research software, research-assistant, predoctoral, and PhD roles in Vienna.
- Scientific Python and research software — refactoring notebooks and scripts into tested packages, command-line tools, and reproducible workflows.
- ML and model evaluation — data alignment, validation design, appropriate metrics, error analysis, uncertainty reporting, and regression tests.
- Data pipelines and quality control — time-series processing, structured metadata, traceable decisions, provenance, and automated reports.
- Debugging and upstream contributions — minimal reproductions, numerical and API-safe fixes, focused tests, CI, and technical documentation.
- NeuroData Release Security Audit — local, read-only checks of privacy-relevant metadata, file coverage, references, and integrity in neurodata release candidates; current prerelease:
v0.2.1b1. - Sleep-EEG staging evaluation — external evaluation of YASA on 20 Sleep-EDF recordings and 28,259 aligned epochs; current release:
v0.3.1with a Zenodo DOI. - Dense-EEG stop-signal pipeline — traceable QC, event reconstruction, reviewed ICA, provenance, and synthetic benchmarking for 129-channel stop-signal EEG; current release:
v0.3.0. - Neural dynamics models — tested simulations of equilibrium potentials, conductance dynamics, spiking networks, and graph topology; current release:
v0.1.0. - OpenSesame visual-world demo — an eight-trial auditory visual-world software demonstration using generated stimuli; current release:
v1.0.0.
Merged
- BIDS schema checks for behavioural files with onset and duration and complete BrainVision file triplets.
- A PyBIDS indexing fix that keeps entity parsing within the dataset root instead of matching entity-like parent directories.
- MNE-Python contributions covering CUDA-backed Hilbert transforms, epoched EEGLAB files without events, OpenBLAS threads, and docstring parameter types.
- MNE-BIDS ecosystem validation for tracking-system metadata, rest epochs, nested BIDS roots, and decoding with too few epochs.
- SleepECG contributions covering external actigraphy inputs, repeated searchback scans in unusable ECG segments, and a CAP Sleep Database reader.
Open
- Parallel manual and automated sleep-stage annotations in BIDS/HED.
- HED schema-manifest cache discovery.
Core: Python, NumPy, pandas, SciPy, pytest, GitHub Actions, numerical validation, time-series analysis, reproducible pipelines, and research software testing.
Scientific domains: MNE, MATLAB/EEGLAB, BIDS/HED, EEG quality control, sleep staging, metadata review, event reconstruction, experimental software, and computational-neuroscience models.