- Published on
Agentic writer: a journey into agentic systems
How I built an agentic writing pipeline on Mastra: researcher, writer, editor, human approval, and the token-limiter bug that made revisions invent details.
Senior Backend Engineer — Security-Focused
Mattia Peretti
I build backend systems that are fast, secure, and built to last — with 10+ years across Node.js and TypeScript — and I'm expanding that into the agentic AI space.
Credentials
- Master’s thesisUsage control monitors — the case of UCLinux
- Bachelor’s thesisRole-Based Access Control (RBAC)
- CertificationEthical Hacker Nanodegree
Selected projects
View all projects →SnapHab
A habit tracking app, in Next.js, that was successfully launched on Product Hunt (#2 product of the week) and garnered positive feedback for its user-friendly design.
Learn more →Agentic writer
A Mastra-powered content pipeline: six AI agents turn notes into researched articles and social campaigns, with a human in the loop before anything goes live.
Learn more →Recent writing
All posts →- Published on
Composer 2.5 vs Grok 4.5 for frontend
Composer 2.5 and Grok 4.5 built the same PancakeBot frontend from the same prompt and skills. Composer finished faster, while Grok delivered the more polished first pass.
- Published on
We just hit the local LLM tipping point
Colibrì, a ~1,300-line pure C inference engine with zero dependencies, runs the 744B-parameter GLM 5.2 MoE model on a laptop with 25 GB of RAM and no GPU by streaming int4-quantized experts from NVMe. Cold-cache throughput is 0.05 to 0.1 tokens per second, so speed is the honest limiter, but it is an existence proof that frontier-scale local inference works on commodity hardware.