I build open-source tools where finance, healthcare, developer infrastructure, and AI systems meet.
My earlier work explored signal processing, knowledge representation, embedded networking, and computer architecture. These days, I’m focused on turning complex domains into practical, understandable software.
-
📊 Copinance — an educational research workspace for equities and options — real market data, inspectable evidence, and the machinery behind the price.
-
🧭 copinance-os — the open research stack behind Copinance: reproducible Python pipelines for equities, options, and macro research, designed to keep AI analysis grounded in market data.
-
🔐 dot-vault — securely back up and restore SSH keys, shell profiles, dotfiles, and terminal configurations across machines.
-
💧 Nefro — a cross-platform hemodialysis companion for patients and nephrologists. Turkish-first, offline-capable, and built around a shared clinical core.
-
🤖 OneMCP — fast, accurate, and cost-efficient API access for AI agents through a unified MCP interface.
- financial research and market structure
- AI agents, tools, and context infrastructure
- healthcare software with real-world constraints
- reproducible data pipelines
- privacy-aware and offline-capable applications
- knowledge representation and graph-based reasoning
The common thread is fairly simple: take a complicated system, find its useful structure, and turn that structure into software people can actually work with.
- 🎭 cox-watermarking — spread-spectrum informed watermarking in MATLAB.
- 🕸️ conceptnet5-client — a Python client and graph-walk inference tools for ConceptNet5, developed during my graduate-school work with the MIT Media Lab.
- 📶 zigduino-aodv-routing — an AODV mesh-routing implementation for Zigduino.
- 🔗 semantic-relation-composition — experiments in relation composition and graph-based reasoning.
- 👁️ Visualize-ConceptNet — a CoffeeScript interface for exploring ConceptNet.
- 🧮 booth-multiplicator — Booth’s multiplication algorithm implemented in MIPS assembly.
- 🎵 music-compiler — a compiler for text-based musical compositions, written in C.
Languages: Python, TypeScript, Java, Scala, C, MATLAB, CoffeeScript, Lisp, and MIPS assembly
Domains: financial systems, AI tooling, clinical applications, knowledge graphs, signal processing, and embedded systems
I’m most at home where research meets implementation—and where “let’s build it ourselves” is still a reasonable answer.
I’m interested in open-source collaboration, particularly around financial research, agent infrastructure, healthcare software, and knowledge-driven systems.
If you’re working on something in that neighborhood, feel free to reach out.