Post 01
Running Firecracker securely
Using the jailer, namespaces and cleanup to run Firecracker safely in production
Writing About AI, Product, And Builder Taste
Notes from building with language models, teams, and small pieces of software.
This is primarily a writing home now: essays, implementation notes, and sharper takes on what feels durable in the current AI wave.
Fig. A • Essays first, projects second, learning all the way through
Fig. 1 - Latest Writing
Fresh Notes
Recent posts
The newest essays, technical notes, and working-through-it posts from the lab.
Post 02
Setting up networking for your VMs
Giving your Firecracker VMs internet access
Post 03
Optimizing for performance
Optimizing Firecracker to run many VMs on the same machine
Post 04
Running a basic VM in Firecracker
The first step towards microVMs
Post 05
A gentle introduction to Firecracker
What is Firecracker and when should I use it ?
Fig. 2 - Projects In Support Of The Writing
Built To Clarify
Selected projects
Small products and open-source experiments that feed back into the essays and technical notes.
Project 01
Flint
Firecracker VM management tool with a Python SDK, daemon API, benchmarks, and optional TUI.
Project 02
Opik Chat History
Extension to log your vibecoding sessions to Opik and make LLM work easier to inspect.
Project 03
AI SDK
Python port of the Vercel AI SDK with a single interface for multiple model providers.