Data Scientist
V. A. Trapeznikov ICS RAS
Jun 2025 — May 2026• Designed relational database architecture on SQLite3; migrated from HDF5 via h5py; structured metadata in JSON – reduced data access time by 80% • Developed production-ready Python script (Pandas, NumPy) for automated time-series reconciliation, saving 5 hours of manual work per week • Performed statistical analysis (Scikit-learn, SciPy) and visualization (Matplotlib, Seaborn), identifying discrepancies in temporal data that informed automated validation pipeline • Implemented XML →HDF5 parsing via h5py, accelerating processing of 9GB video data by 60%