✦ AI Engineering
Engineering the
intelligence layer
for software.
Independent research on AI agents, LLM inference, and retrieval architecture — published as I go, occasionally turned into papers.
╳ Core Research
Current investigations
Two active threads right now: making inference faster and cheaper, and making agent context and retrieval systems that stay reliable as tasks get more complex.
✦ The Piston 1 Logs
Notes from the lab
Research notes, experiments, and breakthroughs in AI engineering and agent systems.
Jun 28, 2026
Inference Pipeline Optimization
Continuous batching and KV-cache reuse — full write-up with benchmarks.
Jun 20, 2026Retrieval Architecture for Agent Context
Vector indices designed for reliable multi-step agent retrieval.
Jun 10, 2026Tool & Context Schema Design Patterns
A taxonomy of tool and context schemas for consistent agent behavior.
May 28, 2026Agent Observability & Tracing
Tracking decision provenance across multi-step agent pipelines.
◆ About
The person behind the lab
Piston 1 Labs is a one-person research and engineering practice — not a euphemism for a bigger team.
Founder & sole engineerI'm the person designing, building, and writing everything at Piston 1 Labs. No team, no agency layer — just one engineer working through inference systems, agent architecture, and retrieval design in public, and taking on a small number of client collaborations along the way.
The Piston 1 Logs are where I publish findings as I go. The longer-term plan is to turn some of this research into formal papers — starting with the work on retrieval architecture for agent context and tool schema design patterns already in the logs.
2026
Started
Solo
Team size
2
Active research tracks
In progress
First research paper
⚡ Collaborate
Selective collaborations
I take on a limited number of projects that align with my research in AI engineering and agent systems.
Currently accepting
AI Engineering & Agent Architecture
Custom inference infrastructure, agent architecture, and retrieval system design for production AI applications. Priority given to projects that advance my research in high‑performance inference and reliable agent systems.
Get in touch