CV

Summary

Ph.D. Candidate in Computational Science specializing in Biosignal Data Engineering, Statistical Modeling, and Automated Analytics Pipelines. Expert in architecting canonical data representations for high-dimensional signals and building versioned, reproducible data products using Python, R, and SQL. Passionate about bridging rigorous mathematical theory—such as Riemannian geometry and domain adaptation—with scalable, production-grade infrastructure.

Education

• Ph.D. in Computational Science (Data Analytics Track), University of Massachusetts Boston, Expected May 2026
• M.S. in Analytics and Modelling, Valparaiso University, May 2018
• B.Eng. in Electrical and Electronics Engineering, Chongqing University, China, Jun 2014

Research & Professional Experience

• Doctoral Researcher (Biosignal Analytics & Infrastructure)

  • University of Massachusetts Boston

    Sep 2019 - Present

  • Multi-Center Data Harmonization: Unified heterogeneous benchmarks (Physionet, BNCI) into a canonical data pipeline. Solved schema inconsistencies across datasets, enabling large-scale cross-study retrospective analysis.
  • Scalable Pipeline Architecture: Optimized parallel workflows on Linux HPC clusters (Slurm). Refactored legacy code into efficient R/Python scripts, reducing cross-validation latency for TB-scale data from days to hours.
  • Data Quality & Robustness: Developed a geometric framework to decouple biological signal drift from system noise. Constructed a “Robustness Atlas” to monitor pipeline sensitivity, serving as a QA layer for analytics outputs.
  • Product-Oriented Model Selection: Operationalized a “Linear-First” deployment rule. Demonstrated that simple, robust linear baselines matched Deep Learning accuracy with 60% less compute, prioritizing maintainability over complexity.
• Instructor of Record (Math Department)

  • University of Massachusetts Boston

    Sep 2023 - Sep 2025

  • Courses Taught: College Algebra (Math 115), Pre-Calculus (Math 130), and Calculus I (Math 140).
  • Role: Designed curriculum, delivered lectures, and managed grading for classes of 30+ students. Synthesized complex quantitative concepts into accessible insights for diverse stakeholders.
• Data Engineer

  • China Mobile IoT Company Limited, Chongqing, China

    Jul 2014 - Aug 2016

  • High-Volume SQL Analytics: Designed and maintained SQL transformations for high-frequency time-series data from millions of IoT sensors.
  • System Reliability: Built dashboards to visualize network anomalies and data flow health, supporting operational decision-making.

Open Source Software & Tooling

• Lead Developer: DA4BCI (R Package)
  • Data Harmonization: Architected a modular R package to standardize Transfer Learning workflows. Implemented canonical transformations (Riemannian Alignment, Optimal Transport) to mitigate distributional shifts.
  • Reproducibility: Built automated testing suites and versioned documentation (CRAN-style) to ensure consistent metric generation.
• Developer: EEG-Feature-Engineering (Pipeline)
  • Automated Processing: Engineered a scalable pipeline to transform raw biosignals into structured feature sets, mapping complex SPD manifolds into vector spaces for SQL-compatible storage.
  • Reliability: Decoupled computation logic from inference APIs to prevent data leakage in production environments.
• Developer: eegwhiten (Data Quality Utility)
  • Developed specialized whitening transforms to normalize high-dimensional covariance matrices, ensuring data quality and statistical validity before modeling.

Technical Skills

• Data Engineering & Pipelines: Automated ETL Workflows, CI/CD Concepts, Data Modeling, Version Control (Git), Docker, Linux/HPC (Slurm).
• Languages: SQL (Complex Transformations), Python (Pandas, NumPy, PyTorch), R (Package Development, Tidyverse), Bash/Shell.
• Biosignal Analytics: EEG Feature Extraction, Canonical Data Representation, Artifact Rejection, Clinical Metadata Integration.
• Mathematics & Modeling: Signal Processing (DSP), Covariance Matrix Estimation, Convex Optimization, Riemannian Geometry, Bayesian Inference.

Awards & Community

• Silver Medal (Top 5%), Kaggle (HMS - Harmful Brain Activity Classification), 2024
• Doctoral Fellowship, University of Massachusetts Boston, 2019 - 2025
• Volunteer Data Scientist, Statistics Without Borders, 2020 - 2021
• SAS Certified, Applied Econometrics and Data Science, 2018

Selected Manuscripts & Presentations

• Shen, Y., et al. Drift-Feature-Performance Decomposition via Structured Geometric Modeling. (Under Review).
• Shen, Y., et al. Confidence-Gated Adaptation for Cross-Session BCI. (Under Review).
• Invited Talk: Benchmarking Classification Pipelines Within and Cross Sessions on the PhysioNet EEG Dataset. MIND Seminar, Inria, France. (Jun 2025).