I'm a PhD student in Computational Science at Florida State University, where I work on agentic artificial intelligence and knowledge graphs for healthcare and biomedical research.
Research
My current focus is agent-based efficiency: developing strategies that let AI agents work against large-scale clinical datasets for fast, cost-efficient data extraction and biomedical knowledge discovery. A central thread of this work is Synapse, a multi-agent knowledge graph construction framework I built on LangGraph and a Neo4j-oriented architecture, designed around evidence-backed schema design, validation, provenance tracking, and graph persistence. Its pipeline spans planner, extractor, validator, schema builder, critic, normalization, and human-review phases, with structured artifact contracts covering evidence spans, citations, schema proposals, confidence vectors, and graph write operations.
Alongside this, I build the infrastructure my research depends on:
- Mango, a local-first autonomous agent platform (Django/LangGraph, Celery/Redis task queues, local llama.cpp inference, PostgreSQL, Uvicorn ASGI) intended as a reusable backend for long-running AI workflows.
- Local LLM infrastructure, including llama.cpp tooling for hardware-aware model serving and a Django-based vLLM control plane with SSH utilities, telemetry management, authenticated operator access, and deployment hardening.
- Agentic tooling, such as AutoNotebooks and Calliope, for generating notebooks, slides, reports, quizzes, and other structured research outputs.
Background
I came to computational science by way of biochemistry, completing a B.S. at Florida State in 2024 with credits equivalent to a second degree in Physical Sciences and Biomathematics, along with minors spanning computer science, mathematics, statistics, physics, astrophysics, and biology.
Before joining PULSE, I was a computer vision researcher in the VIMS Lab at the University of Delaware (2024–2025), where I helped fine-tune a transformer-based CSR model for detecting sickle cell retinopathy and investigated LiDAR-based fall detection for rehabilitation patients. Earlier, as a data analyst and developer in the Lester Lab at Florida State (2023–2024), I worked on data acquisition and modeling for climate challenges in Bonaire and built an interactive visualization dashboard using early versions of Shiny for Python.
My technical work centers on Python for deep learning, C++ for high-performance computing, and SQL for databases, with regular use of PyTorch, MONAI, OpenCV, SimpleITK, scikit-learn, and LangGraph. I also build full-stack research tooling with Django/Flask, Astro, Svelte, and Tailwind/UnoCSS.
Beyond research
Outside the lab I'm a rock climber and PC gamer, and I spend time on worldbuilding and tabletop RPG design — interests that feed directly back into my side projects on knowledge graphs and narrative systems.
