I'm building GraphRAG applications at PozeSCAF Discovery Solutions, using Neo4j knowledge graphs and LLMs to enable natural-language querying across large-scale structured data.
I've turned prototype notebooks into modular, object-oriented Python packages and developed Generative AI workflows that automate domain-specific analysis and improve knowledge retrieval.
Previously, at Immunocure Discovery Solutions, I researched AI/ML trends in drug discovery, prepared data, and trained machine learning and deep learning models for drug property prediction. I experimented with GNN layers and achieved 82% accuracy in property prediction.
I've also authored more than 100 articles on Python and machine learning as a Technical Content Writer, and my work includes projects in attention monitoring, medical-data anomaly detection, and publications on steganography and emotion recognition.
