After spending the last few months head-down in open-source maintainership and production engineering, one thing is abundantly clear:
The line between modular architecture and AI-assisted workflows isn't just blurring - it's disappearing.
As applications scale and dev teams grow, building maintainable web systems requires both robust architectural guardrails and pragmatic, developer-first AI tooling.
Here is what I am actively focusing on and building in 2026:
-
Next-Gen Micro Frontend Architectures 🧩
Continuing to push the boundaries of modular frontend systems. This means advancing Piral, driving forward the up-and-coming Picard.js library, and chairing efforts around Native Federation to keep web applications scalable without build-step bloat.
-
Practical AI Engineering for Developers 🤖
Moving beyond basic LLM chats into real workflow automation. I'm building and testing local workflows with coding agents (Claude, Codex, Copilot) and self-hosted models (Qwen, Gemma) using tooling like Pi and OpenCode to optimize day-to-day maintainer efficiency.
-
High-Performance .NET & Web OSS ⚡
Sustaining and scaling core .NET open-source infrastructure - from the recent 1.0 ecosystem milestones in AngleSharp to cross-platform tooling with Electron.NET, netpack, and Mages.
-
Nanoservices & Lightweight App Design 🏗️
Refining patterns for modular, low-overhead web services that pair seamlessly with decoupled frontends.
Over the coming weeks, I'll be sharing technical breakdowns, code snippets, architecture decision records, and practical workflows around these topics.
If you're working on modular web apps or integrating AI agents into your development stack, I'd love to connect.