Thirshe Perera

youremail@example.com

GitHub · LinkedIn · Twitter

Hi, I'm Thirshe — a second-year Data Science undergraduate at SLIIT, working on computational biology and protein structure prediction. I'm currently designing a hybrid architecture for protein structure prediction, combining efficient long-sequence models with equivariant graph networks, and working toward a research assistant role in computational biology.

Outside research, I work on applied computer vision and machine learning projects, and I'm building toward a long-term venture at the intersection of ML tooling and biology. I'm also involved with SEDS (Space Exploration and Development Society).

Projects

• Protein Structure Prediction Pipeline — novel architecture combining Hyena operators, a temporal encoding bridge, an SNN/LNN layer, and SE(3)-equivariant EGNN layers.
• YOLOv8 PCB Defect Detection — end-to-end computer vision pipeline fine-tuned on a 7-class dataset, achieving mAP50: 0.995.
• DRiST — Diabetes Risk Screening Tool, using Logistic Regression, XGBoost with SHAP, and FT-Transformer models.
• CubeSat NDVI Pipeline — remote sensing pipeline built through SEDS.

Now

rn im doing some awsome projects and research stuff!