Software Engineer and independent researcher exploring the intersection of computational systems and emergent behavior.
My work centers on bringing high‑performance native code into modern web products, translating complex C++ systems into fast, reliable experiences that run directly in the browser using WebAssembly. In addition to the integration work, I implement production features in both the C++ codebase and the TypeScript frontend.
I like working on problems at the boundary of theoretical computer science and practical implementation. At the moment, that translates into an interest in the foundations of modern ML systems: distributed and low‑level infrastructure for training and serving models, retrieval‑augmented generation over vector databases, and reinforcement‑learning‑driven approaches to building adaptive AI tools.
This portfolio documents ongoing research and experimental projects. Approaches are iterative and subject to revision as new insights emerge.
Remote
Product development and research. Currently working on the ingration of high‑performance C++ systems in the browser, with a focus on the performance, safety, and maintainability of the boundary between native code and the web platform.
Remote
Led development of core product features for a B2B SaaS platform, working across the main application stack to ship user‑facing functionality in a production environment. Explored and implemented geospatial analysis capabilities, building early‑stage tools that integrated spatial data and mapping workflows into the product to inform the roadmap.
Web design agency
Led development of open‑source geospatial platforms for climate risk visualization, including the ██████████████████ and ██████████████ , and built interactive data tools for ████████████████ .
Google for Startups, prev. Google Campus
Designed and implemented exploratory tooling to support ████████'s data science platform, including a browser extension and a desktop application for dataset indexing and discovery across distributed research workflows.
Stochastic simulation using a binary grid EBM sampled with wind-aligned blocked Gibbs. View project →
The probability density distribution of stars within a globular cluster. Sampling and visualizing data.
Evaluate and enhance a serverless vector database to better support RAG workloads applied to multimodal datasets.
All project source code and documentation available upon request. Collaborative inquiries welcome.