Shirin Khanam
Machine Learning Research Scientist
Seattle,WA
Summary
Machine Learning Research Scientist and Data Scientist with a Ph.D. in Computer Science and 6+ years of experience developing and deploying scalable machine learning and deep learning systems for scientific and real-world applications. Expertise in PyTorch, NLP, transformer-based models, LLM finetuning, representation learning, and large-scale experimentation on HPC and cloud infrastructure. Proven track record conducting rigorous ML research, building production-oriented AI solutions, and collaborating across multidisciplinary teams to translate advanced research into impactful products and tools.
Overview
16
16
years of professional experience
Work History
Research Scientist 3
University of Washington
Seattle, WA
10.2021 - 04.2026
- Designed, fine-tuned, and evaluated ML models (logistic regression, random forest, neural networks, topic models) for disease diagnosis and grouping, achieving up to 90% classification accuracy on sensitive biomedical datasets.
- Led end-to-end analytical workflows, including data curation, feature engineering, rigorous experimental design, model selection, validation, interpretation, and reproducible pipeline development.
- Conducted robustness, feature importance, and bias analysis using cross-validation, data augmentation, sensitivity testing, ablation studies, and SHAP-based interpretability to ensure model reliability and generalizability across cohorts.
- Served as technical lead across 10+ clinical and biomedical studies, integrating NLP pipelines with statistical inference to identify data inconsistencies and support successful grant funding.
- Partnered with interdisciplinary research teams including clinicians, bioinformaticians, and data scientists to translate complex biomedical research questions into scalable ML solutions.
- Architected and deployed scalable ML systems using AWS Lambda, API Gateway, and SLURM-based HPC job orchestration for parallel and sequential workloads, reducing end-to-end pipeline runtime by ~50%.
Graduate Researcher
La Trobe University
Melbourne, Australia
08.2016 - 06.2021
- Built and optimized deep learning and NLP-based text summarization models using Transformers, LSTM/RNN, CNN-attention, Word2Vec, and BERT.
- Applied supervised and unsupervised machine learning techniques (SVM, Random Forest, LDA, NMF) for classification, clustering, and pattern detection.
- Improved model reliability through fine-tuning, reinforcement learning, and regularization; contributed to peer-reviewed publications including IJCNN 2021.
- Performed data extraction, preprocessing, integration, and analysis for large-scale NLP datasets, and engineered end-to-end ETL and SQL-based data transformation pipelines to support feature engineering and model development.
Research Assistant
Sungkyunkwan University
South Korea
03.2013 - 02.2015
- Developed graph-based dependency and similarity algorithms using ontologies and bipartite networks.
Sr. Software Engineer
Rich Business System Ltd.
Bangladesh
02.2010 - 12.2012
- Built and maintained enterprise backend systems supporting 1M+ records.
- Optimized SQL queries, indexing strategies, and data workflows.
Education
Ph.D. - Computer Science & Information Science
La Trobe University
01.2022
M.Sc. - Electrical & Computer Engineering
Sungkyunkwan University
01.2015
B.Sc. - Computer Science & Engineering
CUET
01.2007
Skills
- Machine Learning & AI: Supervised/unsupervised ML
- Deep learning
- NLP
- LLM fine-tuning
- Programming: Python
- PyTorch
- Statistical Analysis
- Data Management: ETL pipelines
- Large-scale data cleaning
- SQL optimization
- Cloud & HPC: AWS Lambda
- API Gateway
- Docker
- SLURM
- Model Evaluation & Interpretability: SHAP
- Calibration
- Robustness analysis
Accomplishments
- LumpIt (Predictive Tool), End-to-end ML predictive modeling, AHSG Consortium 2024, and API deployment for rare disease classification using AWS Lambda and API Gateway- Website for prediction: https://depts.washington.edu/jxchong-lab/LumpIt/
- Khanam, S., et al., "Concept-based Topic Attention for Convolutional Sequence Summarization." International Joint Conference on Neural Networks 2021. Paper
- Winner, Best open-source ML solution for rare disease research: ETL, phenotype encoding, NLP/embedding workflows - https://github.com/jxchong-lab/GlobalGenes_RareX_Challenge_2023
Personal Information
- Citizenship: US Citizen
- Visa Status: US Citizen
Timeline
Research Scientist 3
University of Washington
10.2021 - 04.2026
Graduate Researcher
La Trobe University
08.2016 - 06.2021
Research Assistant
Sungkyunkwan University
03.2013 - 02.2015
Sr. Software Engineer
Rich Business System Ltd.
02.2010 - 12.2012
Ph.D. - Computer Science & Information Science
La Trobe University
M.Sc. - Electrical & Computer Engineering
Sungkyunkwan University
B.Sc. - Computer Science & Engineering
CUET