Systems
Backend and infrastructure work
I gravitate toward reliability, delivery pipelines, platform engineering, and performance tuning problems, especially when messy systems can be turned into something cleaner and easier to operate.
Software engineer, product experimenter, public writer
I turn systems, tools, and ideas into shippable work.
I focus on backend systems, cloud-native infrastructure, and small but complete product experiments. This page is meant to work as a portfolio front door first, then a blog: projects, writing, and the kinds of problems I like solving live together here.
Backend Systems
Cloud Native
Product Experiments
Technical Writing
Products
Small products with clear edges
I keep building compact tools that can actually be shipped and used, with an emphasis on clarity, constrained scope, and interfaces that feel intentional.
Writing
Public notes that show the work
The writing archive is mostly about debugging, architecture tradeoffs, and practical infrastructure detail. It doubles as both blog and proof of work.
Selected work
Projects that represent how I build
Live product
Track3
Portfolio analytics for centralized exchanges and Web3 wallets, built around local encryption and zero-backend ownership of data.
Desktop experienceTuneBox
A retro-styled desktop music player where interface design, lyric handling, and tactile controls matter as much as the core playback flow.
Utility toolETF Overlap Calculator
A focused finance utility that makes ETF overlap and unintended concentration easier to inspect without burying the user in raw holdings data.
Recent writing
Latest posts
My Daily AI Workflow: Cutting $150/Month with the Right Model Mix
How I combine multiple AI products to boost efficiency and cut costs
I Gave Up on Obsidian and Notion and Built My Own AI-Powered Second Brain
I gave up on Obsidian and Notion and built my own AI-powered second brain
Best Practices for Go Monorepo Dependency Management
Best practices for Go Monorepo dependency management, using modular go.mod files and replace directives to solve version conflicts and improve build efficiency.
ArgoCD Monorepo Performance Optimization (Part 2): Cache, Sidecar & Git Tuning
Part 2: How we reduced ArgoCD sync time from 3min to under 30s — cluster cache tuning (ARGOCD_CLUSTER_CACHE_RESYNC_DURATION), sidecar plugin tarball optimization (95% CPU drop, 3s → 400ms), proactive git fetch, and scheduled git gc with readiness probe gating.