Netsanet Tesfaye

At iCog Labs, I contribute to AI research and development focused on symbolic reasoning, probabilistic inference, and knowledge-driven systems. I work with OpenCog Hyperon-inspired techniques and weighted inference for explainable AI applications.

For Rejuve, I developed and optimized subgraph mining algorithms for biomedical knowledge graphs. I used graph-theoretic pruning, indexing, and probabilistic scoring to improve pattern detection across heterogeneous datasets.

I built an end-to-end plant disease detection pipeline using 86,000+ labeled images, achieving 94% top-1 validation accuracy with transfer learning and a VGG-style CNN. I deployed the inference service with FastAPI and a React frontend.

I also developed a persistent-memory email agent with LangChain and LangGraph, and designed real-time embedded control systems. Through SheCodes, I mentor aspiring women developers in HTML, CSS, JavaScript, and React.