Statistical Machine Learning · Model Evaluation · EEG & BCI · Research Software
I study how models perform when data changes. My work evaluates models using the history available at deployment, selects useful training data, and measures future gains alongside individual performance loss after updating. EEG and brain-computer interfaces are my main research setting. I build Python and R tools for reproducible experiments and result inspection. I completed my Ph.D. in Computational Sciences at UMass Boston in August 2026.
Selected Evidence
Cross-Session EEG Transfer
Confidence-Gated Source Selection Under Session Drift Matched evaluation of global and selective source use · Published Aug 2026Research Software
CrossDA · ShiftDx · NeuroStream Transfer · drift diagnostics · pseudo-online replayEEG Competition
Kaggle HMS Silver Medal Rank 98 / 2,767 · Top 4%Selected Research
Published · Frontiers in Human Neuroscience - Brain-Computer Interfaces · 28 Aug 2026
Matched evaluation of full pooling, distance weighting, and confidence-gated source selection for cross-session motor-imagery EEG transfer. doi:10.3389/fnhum.2026.1895016
Under Review · IEEE Access
How evaluation changes when a decoder can train only on recordings available before the test session.
Working Paper
Low-channel EEG adaptation can improve average future classification while individual participants lose performance on earlier test data.
Recent preprint: Covariance-Aware Augmentation for Cross-Session BCI: A Multi-Dataset Evaluation · Research summary
More research: Source screening · working paper · History replay · Calibration-cohort selection · Dissertation
Research direction in planning: Sparse EEG for Movement and Motor Learning
Software & Projects
Understanding model degradation and adaptation
Offline analysis of model degradation, adaptation benefit, and retraining gaps across EEG sessions, with fixed-reference comparisons and statistical reports.
Role: Dashboard developer
Evidence: Five drift metrics; fixed-reference analysis; no-adaptation, adaptation, and retraining comparisons
Reproducible evaluation of transfer-learning pipelines
CLI and Python API for comparing transfer-learning strategies with matched experiments, detailed results, and auditable session roles. Evaluated using repeated-session EEG.
Role: Python package developer
Evidence: CLI + Python API; MAP, DWP, MMP, and BDP; auditable summary, detail, and session-role outputs
Sensor placement optimization and scientific simulation
An interactive application for electrode-layout design, bounded optimization, and forward simulation. A 3D interface connects to Python computations and reusable local results; evaluations use one simulated head model.
Role: Research software development
Evidence: Constrained layout search; field reconstruction; traceable computation records
Teaching & Student Research Mentoring
I taught Precalculus (MATH 130) and Calculus I (MATH 140) as an instructor of record at UMass Boston across multiple semesters in 2022–2024. My responsibilities included preparing and delivering lectures, assigning homework, designing assessments, grading, and exam preparation.
I also have mathematics and statistics teaching-assistant experience at UMass Boston and Valparaiso University. In the 2025–2026 academic year, I co-mentored an undergraduate EEG classification project through the STEMPOWER fellowship, supporting benchmark replication and a campus poster.
Technical Notes
How to fit feature extraction on training data and reuse the fitted transform for test trials.
The eigendecomposition and variance scaling behind PCA whitening, with the corresponding R implementation.
How alignment can help transfer, produce negative transfer, and be checked against a no-adaptation baseline.
Experience
Sep 2019 - Aug 2026
Cross-session EEG transfer, drift diagnostics, statistical validation, and reproducible Python/R research software.
Jul 2014 - Aug 2016
Time-series sensor data, MySQL performance, monitoring dashboards, and automated reporting.
Contact
I am interested in academic and research positions in statistics, machine learning, and scientific computing, including roles that combine teaching and research. I also welcome research collaborations in EEG/BCI, model evaluation, and scientific software.