I recently graduated with a Master's degree in Computational Biology from Carnegie Mellon University (Class of 2026), bringing 3+ years of industry experience as a bioinformatics engineer. My long-term vision is to advance personalized, preventive medicine through predictive, in silico modeling and scalable AI.
I specialize in bridging biological research and software engineering—focusing on architecting data pipelines, training specialized AI models, and deploying production-grade MLOps infrastructure for complex biological datasets.
-
An iOS and watchOS application designed for on-device panic attack monitoring, utilizing agentic workflows and physiological features.
-
A distributed active learning framework for antigen-antibody binding prediction using Protein Language Models (PLMs).
-
A demonstration of a fine-tuned language model for classifying patient symptoms, served via a FastAPI/Redis-backed REST API.
At CMU's Safe AI Lab, I developed a Mamba-based Foundation Model for clinical time-series data. By leveraging Self-Supervised Learning (SSL), the model is designed to capture long-range dependencies in physiological signals (ECG) with linear complexity.
Key Achievement: Successfully benchmarked robust length generalization for a model trained on 10-second segments. It demonstrated zero-shot robustness on shortened 2-second inputs (AUROC drop <1%) and seamlessly scaled to unseen 30-minute external records, achieving >3x higher inference throughput compared to baseline architectures.
- Languages: Python, SQL, R, Bash
- Frameworks & Libraries: PyTorch, Hugging Face (Transformers, PEFT), Ray, XGBoost, LightGBM, CatBoost, Scikit-learn, Pandas, NumPy, OpenCV, Matplotlib, ggplot2, Seurat, Scanpy
- AI & ML Techniques: Self-Supervised Learning (SSL), Active Learning, LLM Fine-tuning (LoRA), Prompt Engineering (CoT), Feature Engineering
- MLOps & Backend: AWS (SageMaker AI, S3, ECS/Fargate, ECR, DynamoDB), Docker, Git, GitHub Actions (CI/CD), MLflow, Wandb, FastAPI, Redis
- LinkedIn: URL