Projects that ship
Agentic AI platforms, product backends, open-source infrastructure, and published research. Mostly production systems with real users.
Tobu
AI-powered personal memory platform, live on Google Play
Developed and maintained the backend infrastructure, with integrated LLMs, and took the system to production fast enough to support the company's $1.5M raise. Shipped ML features for image analysis and conversation processing that turn unstructured photos and chats into structured, searchable information, and built PySpark pipelines for user analysis and feature optimization.
1,000+ users · $1.5M raiseJaseci Stack
Open-source core contributor (jac-scale, jaclang)
Core developer of jac-scale, Jac's Kubernetes deployment and scaling layer that takes identical local Jac code to production, auto-provisioning databases, Redis caching, and authentication with zero configuration. Contributor to jaclang, an AI-native language featuring graph-based Object-Spatial Programming with native Python, JavaScript, and C interoperability.
70+ merged PRs · 950+ GitHub starsFlowline
AI-powered project management platform
Lead developer of a Kubernetes-deployed platform that syncs with GitHub and keeps the team board current as issues and pull requests change. Its AI assistant answers questions about the team's work, writes daily standup notes, and turns plain-language requests into tasks.
LiveSpenzo
Personal money manager
Lead backend developer of the Go API behind the mobile and web apps for tracking income, expenses, budgets, and net worth, with an OpenAPI contract, contract tests, and CI/CD from GitHub Actions to Google Cloud Run and Cloud SQL.
Live on the App StoreMesh
Realtime team chat
Main developer of a Slack-style chat product with web, mobile, and desktop clients. A Go modular-monolith backend and WebSocket hub use per-channel sequence numbers to drive message ordering, unread state, and reconnect backfill.
LiveHabitat
A secure home for AI agents on computers and phones
An open-source design for giving an agent its own persistent space. The agent will run inside WebAssembly with no outside access, and a trusted Rust broker will mediate every request to reach the network, files, models, or secrets. Currently at the design and project skeleton stage.
Early developmentPersonal project
Alfred
A personal AI desktop agent
An AI agent designed to run natively on macOS, Windows, and Linux. You will talk or type to it, and it will get things done on your computer with your permission, through a Rust policy layer the model can't influence. Currently in its foundation phase.
Early developmentIndustry
Einstein Pro
AI patent generation & document management
Backend developer for an AI platform that generates patents and manages documents, including a RAG-based chatbot over a vector database for fast, grounded answers.
Self-Adaptive NILM
Deep learning research → 2 IEEE publications
Designed WaveNet/CNN models with ensemble methods for appliance-level power prediction; applied pseudo-labeling and transfer learning for self-adaptation to appliance aging without retraining, reaching over 97% accuracy on a six-year dataset. Published at IEEE World AI IoT Congress 2024 and in IEEE Access.
Best Presented Paper · IEEE AIIoT 2024