Home

Yunfei (Eric) Liao

Senior ML/Data Engineer

Senior ML/Data Engineer with a Ph.D. in Bioinformatics and a background in statistical modeling, feature engineering, and large-scale data analysis. I build production data and machine learning systems, agentic LLM applications, and deployed tools that turn raw data into useful outcomes.

Eric Liao

📧 Email: [email protected]

📞 Phone: +1 704-340-5107

📍 Location: Charlotte, NC, US

📪 LinkedIn: linkedin.com/in/ericlyf

🖥️ GitHub: github.com/Afei99357

🌐 Website: liaoyunfei.name

🔬 Google Scholar: Yunfei Liao

🎓 Education

  • Ph.D. in Bioinformatics
    University of North Carolina at Charlotte (UNCC)
    2020–2025

  • M.S. in Information Technology
    University of North Carolina at Charlotte (UNCC)
    December 2019

  • M.Eng. in Materials Processing
    Nanchang Hangkong University (NCHU), China
    June 2016

  • B.Eng. in Materials Processing
    Nanchang Hangkong University (NCHU), China
    June 2013

🧳 Professional Experience

  • Senior ML Engineer
    Blueprint Technology
    June 2025 – Present

    • Architected an agentic LLM platform that reads legacy NiFi flows and generates validated Databricks notebooks for Hadoop-to-Databricks migrations.
    • Built an internal flow-analysis application with dependency graphs and automated validation, used daily by 30+ engineers.
    • Built a production call-intelligence and churn-detection pipeline for 35k+ calls per year using GPU transcription, speaker diarization, LLM summarization, and ML classification.
    • Created a Databricks App for exploring call insights, churn drivers, and sentiment trends.
    • Technologies: Python, PySpark, SQL, NetworkX, Databricks, AWS, Hadoop, NiFi, Docker, Whisper, Llama, MLflow, Streamlit

  • Research Assistant
    Elizabeth Cooper Lab - UNCC
    January 2023 – May 2025

    • Led interdisciplinary research on West Nile Virus (WNV) and population genetics of Culex mosquitoes.
    • Processed 100+ GB climate and ecological datasets on a 16 GB laptop using memory-mapped and streaming workflows.
    • Built automated scraping, image-processing, orchestration, and containerized RAG pipelines for public-health surveillance data.
    • Technologies: Python, Airflow, DBT, DuckDB, FastAPI, Selenium, OpenCV, scikit-learn, MLflow, Docker

  • Research Assistant
    Xiuxia Du Lab - UNCC
    May 2019 – December 2022

    • Designed and implemented machine learning algorithms for metabolomics-based biomarker discovery.
    • Reduced large-scale mass-spectrometry comparison from quadratic to approximately log-linear complexity.
    • Technologies: Python, scikit-learn, MySQL, JavaScript, HTML/CSS

  • Project Manager
    AVIC Digital
    June 2016 – July 2018

    • Managed Manufacturing Execution System (MES) projects for aviation manufacturing.
    • Translated customer requirements into actionable technical designs with a 100% project completion rate.

⚙️ Projects

Agentic Databricks Documentation Assistant

  • Built a grounded RAG application over official Databricks documentation, with an agentic retrieval loop and clickable source citations.
  • Implemented incremental source refresh, chunk and embedding reuse, FAISS snapshot validation, and local SQLite or governed Databricks deployment modes.
  • Evaluated tool-calling reliability across local and hosted models; Claude Sonnet 4.5 was the most dependable model for the complete multi-step retrieval loop.

West Nile Virus Prediction at the Cooper Lab

  • California WNV data visualization illustration (click)

  • Sole investigator for a profoundly interdisciplinary project analyzing WNV antecedents, overturning many long held expectations and highlighting new avenues for investigation

  • Collected meteorological, demographic, ecological, genetic and CDC disease surveillance data from over 1000 sources, ranging from gigantic (Copernicus) to arcane (CA Arbovirus Bulletin)

  • Interpolated, resampled, repaired, and consulted relevant govt authorities on data irregularities

  • Collaborated with USDA, CDC, CA Public Health Dept, taking an active role in data acquisition

  • Exhaustively analyzed datasets with a wide range of discipline-specific statistical algorithms

  • Innovatively exposed sensitivities and relationships using a wide variety of statistical analyses

  • Automated dozens of processes, ranging from scraping (beautifulsoup, selenium), to computer vision (cv2), even orchestration and database ingestion (Airflow)

ELT Pipeline Extracting Climate Data at the Cooper Lab

  • Designed and implemented an automated ELT pipeline to extract, load, and transform daily weather data from the Open-Meteo API using Apache Airflow, DBT, DuckDB and AWS.
  • Orchestrated data workflows in Airflow, triggering DBT runs via Dockerized environments to ensure seamless data processing.
  • Developed DBT models to transform and structure climate data for efficient analysis and visualization.
  • Scheduled daily updates, integrating Airflow DAG execution with a build evidence project, then pushed processed data to AWS S3 for cloud storage and further analysis.
  • Built an interactive climate visualization that enables users to dynamically explore climate trends across the U.S..

Population Genetics in Culex Mosquitoes at the Cooper Lab

  • Conducted population genetics research on Culex tarsalis to identify genetic adaptations to environmental factors such as temperature and precipitation
  • Collected a wealth of environmental data similar to the WNV project mentioned above
  • Used field-specific statistical models like Redundancy Analysis to search for significant environmental correlations and effects on adaptiveness of individual SNPs
  • Displayed discovered connections using a variety of visualizations such as upset plots and pairwise plots of several kinds

Biomarker Discovery at the Du Lab

  • Designed and implemented supervised machine learning algorithm to detect early disease via un-targeted metabolomics studies using scikit-learn

Automated Data Analysis Pipeline at the Du Lab

  • Developed a memory-efficient prescreening algorithm for mass spectrometry search
  • Created visualizations for large datasets using html and javascript

MES Project Management at AVIC Digital

  • Led projects focused on Manufacturing Execution System (MES) for an airplane manufacturer, ensuring seamless integration of software with manufacturing workflows
  • Gathered customer’s application requirements and translated into actionable designs for development teams

📊 Skills

  • LLM & GenAI: Agentic LLM systems, RAG, vector databases (ChromaDB, DuckDB), prompt engineering, model quantization, Llama, Whisper, MLflow
  • ML & Statistics: scikit-learn, statistical modeling, feature engineering, time-series analysis, classification, clustering, model evaluation
  • Data & Infrastructure: PySpark, Databricks, Airflow, NiFi, Hadoop, DBT, Docker, AWS (S3, EC2)
  • Programming Languages: Python, SQL, R, JavaScript
  • Spoken Languages: English (Fluent), Chinese (Native)

🏅 Certifications

  • Databricks Certified Generative AI Engineer Associate — Databricks
  • Databricks Certified Data Engineer Professional — Databricks

📎 Publications (* co-first author)