Projects

Built to
ship. Built to scale.

Hackathon wins, production AI, open source I actually use. Every project here started with a real problem, and ended with something that works.

Personal Projects

dekko

Personal

A fast, offline, dependency-free static code map generator and codebase indexer for LLM coding agents. Parses a repo once with tree-sitter into a call graph, cutting token usage 3x-200x versus a Read/Grep workflow across benchmarked real-world repos.

Open source, live on PyPI

Python Tree-sitter MCP CLI

Claude Castor

Personal

An MCP server that offloads Claude Code's large-context work, file exploration, indexing, summarization, deep research, to Google's free Antigravity CLI. Claude orchestrates, Antigravity does the heavy lifting, and it runs with no API key required.

Open source

Python MCP FastMCP Antigravity

Printify Python Client

Personal

A clean, typed Python client for the Printify API. I built it for myself, then figured the community could use it too, so I published it to PyPI.

Live on PyPI

Python API Development REST

Zed: Copy Markdown as Rich Text

Personal

An open source contribution to Zed adding rich-text clipboard support to the Markdown Preview. Copying a selection now writes both plain text and HTML, so rich-text targets like Notion or Google Docs get formatted output instead of raw markdown, across macOS, Linux (X11 + Wayland), and Windows.

Open source contribution

Rust GPUI pulldown-cmark

Professional Projects

GenAI Document Analysis Platform

Professional

Ally's first production GenAI, and the one I'm proudest of. End-to-end document analysis with natural language Q&A on proprietary data. I authored the Model Development Documentation and cleared the full Model Risk Management (MRM) tollgating myself. That made it the governance-approved template every AI initiative since has followed.

$300k annual savings, enable high volume competitor analysis
Python LangChain RDS Vector OpenAI RAG Next.js

Call Center Transformation App

Professional

Customer data was scattered across five systems, and reps were digging through all of them mid-call. I pulled it into one platform. Hours saved on every call, and a team that finally stopped fighting their tools.

$1.1M ROI, 60% faster call handling
Full Stack Data Integration UI/UX

Automated Terraform Updater

Professional

2025 hackathon finalist, picked up for production. A GenAI tool that hunts down non-compliant Terraform modules across GitLab repos and fixes them before they ever become a problem.

$100k+ projected annual savings, selected for production development
Python OpenAI Terraform GitLab API

AI Code Review Pipeline

Professional

A CI/CD pipeline that puts Claude on every commit: writing unit tests, reviewing the code, and blocking the bad stuff before it ever merges. Org-wide.

Enforced TDD across engineering team
OpenAI CI/CD Python Git Actions

Automation Development Kit

Professional

A Python library the whole org runs on. One home for logging, metrics, database integrations, and API clients, so nobody has to reinvent the wheel on every new project.

Standardized development across automation teams
Python Snowflake Streamlit Nucleus Nexus Oracle

Got a problem worth solving?

I'm always looking for the next challenge. Let's build something that matters.