Matchstick Replay
MaintainedA market-data replay tool — capture a live L2 crypto order-book feed, then scrub, seek, and replay it deterministically, reconstructing the same market state at the same timestamp every time.
Visit →Documentation & AI-Ready Knowledge Systems · Building Matchstick Replay
I help customers succeed with complex technical products through support, implementation, and customer engineering. I started in frontline support at NinjaTrader and moved up to support lead and product manager, then managed technical services at Ookla and worked as a customer engineer at Privacy Dynamics. I'm comfortable moving between customers and engineers, working with logs and traces, and turning recurring cases into better tooling, documentation, and product follow-up.
Support workflows, escalation paths, service knowledge, and team enablement.
Customer setup, migration, technical troubleshooting, and adoption.
Information architecture, onboarding, docs-as-code, and API clarity.
Turning repeated confusion into a workflow people can follow and improve.
Schema validation, CLI tools, and small utilities for documentation-heavy systems.
Customer discovery, product scope, cross-functional priorities, and delivery.
Became the primary at-home caregiver while maintaining technical skills through limited consulting and independent software projects.
Supported clients through dedicated channels and investigated issues across APIs, HTTP requests, logs, browser tooling and SQL-backed workflows. Built small tools and documentation systems to make recurring implementation and support work easier to diagnose and hand off.
Built a documentation website, supported early-stage customers, troubleshot UX/QA issues.
Consolidated product and engineering knowledge into structured wiki content for an Office 365 support organization; collaborated with PMs and engineers and tracked work through Git and Markdown.
Improved workforce-management efficiency by about 10% through process analysis; documented workflows and built SharePoint knowledge hubs.
Managed a 5-person technical services team supporting ~1,500 accounts; reduced customer response time by about 50% and helped customers migrate from Flash-based testing to browser-native WebSocket/XHR.
Led a support team of 15; maintained docs for ~40,000 users; contributed to C# debugging; translated customer workflow problems into product and documentation improvements.
A market-data replay tool — capture a live L2 crypto order-book feed, then scrub, seek, and replay it deterministically, reconstructing the same market state at the same timestamp every time.
Visit →RAG preprocessing library — chunks UTF-8 text and verifies every chunk's byte range and SHA-256 hash against the source. Benchmarked against LangChain (byte-offset drift on non-ASCII text) and on 15 CUAD contracts (100% stale-chunk detection vs. 0% baseline).
Visit →
A free, ad-free PWA that times Japanese interval walking — voice cues, chimes, and a streak-tracking progress screen.
Read more →Schema validation toolkit for content-rich systems and documentation pipelines.
Visit →Dependency-free TF-IDF similarity toolkit. Works in Node, browsers, and Obsidian.
Visit →Reach out about project-based, outcome-scoped work or to talk shop.
Get in touch →