CV
Data Scientist and Researcher with 6+ years of experience in Bayesian modeling, machine learning, and NLP applied to marketing, healthcare and public health.
General information
| Name | Hashan Fernando |
| Title | Data Scientist & Researcher |
| hashan.dan@gmail.com | |
| Phone | (865) 321-5237 |
| Url | https://hashansl.github.io/ |
| Summary | Data Scientist and Researcher with 6+ years of experience applying Bayesian modeling, machine learning, and NLP to marketing, healthcare and public health problems. Proficient in R, Python, and high-performance computing environments. Experienced in extracting real-world clinical data, building reproducible pipelines, and translating technical research into practical solutions. Strong background in applied statistics, deep learning, and large language models. |
Work
-
2022.08 - 2026.05 Oak Ridge, United States
Graduate Research Assistant
Advanced Computing in Health Sciences Section, Oak Ridge National Laboratory
- Extracted, cleaned, and integrated VA electronic health record and mortality data, mapping ICD diagnosis and cause-of-death codes to standardized risk categories to build research-ready datasets for county-level opioid overdose modeling.
- Designed a novel algorithm integrating topological data analysis and Bayesian spatial modeling to identify complex spatial patterns in opioid overdose risk across the USA, reducing uncertainty in baseline Bayesian spatial model estimates.
- Built custom multi-channel CNN architectures in PyTorch to classify regional health risk strata across the USA using high-dimensional topological persistence images.
-
2020.06 - 2022.02 Sri Lanka (Remote)
Marketing Analytics Consultant
Freelance (Fiverr & Upwork)
- Designed and analyzed marketing performance and sales dashboards (React, Google Data Studio) pulling multi-channel spend data via Supermetrics, Shopify, and the Facebook Ads API to identify trends and inform client optimization decisions.
- Managed client relationships and reporting cycles independently.
-
2019.12 - 2022.02 Peradeniya, Sri Lanka
Researcher
Department of Statistics and Computer Science, University of Peradeniya
- Developed a smart seal and audio detection system for protecting steel cargo from tampering, leveraging IoT sensors for real-time monitoring.
-
2019.01 - 2019.07 Colombo, Sri Lanka
Data Scientist Intern
VizuaMatix
- Engineered a machine learning model to identify non-performing customers in a banking system.
- Developed a route optimization pipeline for food distribution trucks to improve efficiency and reduce operational costs.
-
2018.10 - 2019.01 Biyagama, Sri Lanka
Education
M.S. Statistics
2022 – 2026 Haslam College of Business, University of Tennessee, Knoxville · Knoxville, TN GPA: 4.00 / 4.00
Ph.D-Dropout Data Science and Engineering (DSE)
2022 – 2026 Bredesen Center, University of Tennessee, Knoxville · Knoxville, TN GPA: 4.00 / 4.00
M.S. Data Science
2020 – 2022 Postgraduate Institute of Science, University of Peradeniya · Sri Lanka GPA: 3.75 / 4.00
B.Sc. Statistics and Operations Research
2015 – 2018 Faculty of Science, University of Peradeniya · Sri Lanka
Awards
- 2022
Merit Award for M.S. in Data Science
University of Peradeniya
- 2017
Half Colors
Colors Awarding Ceremony, University of Peradeniya, Sri Lanka
Publications
Manuscript in preparation
H. Fernando. Manuscript in preparation, 2026.
BMC Public Health 25, no. 1 (2025): 1759
A. Deas, H. Fernando, A. Heidi, A. Kapadia, J. Trafton, A. Spannaus, V. Maroulas.
Conference Proceedings for International Population Data Linkage Conference, vol. 9, no. 5
H. Fernando, A. Spannaus, A. Kapadia.
Presentations
- 2025.02
Unmeasured Spatial Risk in Disease Modeling: A Topological Data Analysis Approach
UT-Oak Ridge Innovation Institute's Second Annual Research Symposium, Knoxville, TN
- 2024.09
A Geospatial Study of Opioid Overdose Risk in Washington D.C. Metropolitan Area
Interdisciplinary Association for Population Health Science Conference, St. Louis, MO
- 2024.09
A Geospatial Study of Opioid Overdose Risk in Diverse Regions
International Population Data Linkage Network Conference, Chicago, IL
Skills
Languages
R Python SQL Spark C/C++ Java
Statistics & Methods
Marketing Mix Modeling Causal Inference Regression Bayesian Hierarchical & Spatial Modeling Mixed-Effects Models Machine Learning MCMC Deep Learning NLP LLMs Operations Research Topological Data Analysis
Data & Databases
SQL & Relational Databases Data Extraction / ETL Data Cleaning & QA REST / Web APIs Spark
Dashboards & Visualization
Shiny Tableau Looker (Data Studio) React
Frameworks & Tools
PyTorch TensorFlow scikit-learn RStudio Git GCP AWS Slurm Docker PySpark
Languages
| English | |
| Fluent |
| Sinhala | |
| Native speaker |
Interests
Marketing Mix Modeling Causal Inference Bayesian Modeling Machine Learning & Deep Learning Spatial Statistics Topological Data Analysis NLP & Large Language Models Healthcare & Public Health High-Performance Computing
Selected projects
-
1st Place Winner — MOSSAIC Hackathon (ORNL, National Cancer Institute & Leidos Biomedical)
Co-led a research team and empirically demonstrated data privacy leakage in cross-silo Federated Learning architectures by deploying gradient-based adversarial attacks against text transformer models.
-
Clinical Document Classification & NLP Clustering (COSC-526: Data Mining & Analytics)
Developed an end-to-end NLP pipeline on unstructured clinical notes; generated contextual embeddings for supervised classification and applied NER, Mapper-based TDA clustering, cosine similarity, and t-SNE/UMAP for medical terminology extraction and visualization.
-
Spatial-TDA
Developed an open-source Python package designed for extracting topological information from spatial data.
-
Causal Inference for Marketing Analytics (BZAN-593: Independent Study)
Developed a Bayesian marketing mix model to estimate causal ROI of marketing spend from observational time-series data.