Senior Data Scientist & Computational Mathematician
I translate complex mathematics into production software — specializing in topological data analysis, deep neural network explainability, and agentic AI systems, with research collaborations at NASA and DARPA.
Gabrielle Angeloro McKasy
About me
I'm a Senior Data Scientist and Computational Mathematician at Geometric Data Analytics, where I specialize in turning advanced mathematics into deployable software. My core focus is topological data analysis (TDA) and the explainability and trustworthiness of deep neural networks.
I've led research with NASA on DNN reliability in safety-critical systems, contributed to DARPA-funded agentic AI tools for naval design, and collaborated with Lawrence Berkeley National Lab on real-time anomaly detection for synchrotron experiments.
I hold an MS in Mathematics from Iowa State University and a BS in Mathematics, summa cum laude, from SUNY Geneseo. I've published at NeurIPS and IEEE Aerospace, and contributed to the open-source Scikit-TDA ecosystem.
Selected work
Geometric Data Analytics — Durham, NC — Jan
2021–Present
Led a DARPA-funded effort to build an agentic AI assistant with natural language querying, implementing set-based design principles within a modular function-calling framework.
Embedded topological shape invariants into validation and verification systems for deep neural networks, enabling continuous model health monitoring in safety-critical applications.
Co-led anomaly detection R&D for Lawrence Berkeley National Lab's synchrotron experiments, surfacing emerging issues in continuously updating high-throughput sensor data.
Developed and implemented supply chain optimization algorithms for an Army-funded platform, collaborating across frontend and backend teams to integrate into a production schema.
Designed and implemented pathfinding algorithms for aerial vehicle search and rescue operations, delivering efficient route solutions under real-world constraints.
Created and implemented novel topological featurizations for CNN pipelines used in object detection, in collaboration with Raytheon research partners.
Education
Skills
My work sits at the intersection of mathematics, machine learning, and production software. I specialize in applying topological and geometric methods to real-world AI problems — from model explainability to anomaly detection — and I build the software to deploy those ideas at scale.
Contact
I'm always open to interesting conversations — whether you want to collaborate on research, discuss AI safety, or just say hello.
