Inertia Labs is an independent R&D space led by Andy Bell. I build and document practical systems at the intersection of AI agent architectures, Rust, and Web3 primitives (coordination, identity, trust), with continued interest in verifiable computation / ZK where it meaningfully improves system guarantees.
AI agents + coordination: designing practical agent systems where blockchains are used only for what they’re good at (coordination, identity, trust, settlement) — not for compute.
- Agent systems in Rust (tool use, memory, evaluation, reliability)
- Retrieval & knowledge systems (RAG as a foundational agent capability)
- Agent identity, trust, and provenance
- Coordination patterns (task markets, verification, dispute handling; hybrid on/off-chain architectures)
- Verifiable computation / ZK (selective use for auditability and guarantees, not novelty)
Most repos are “lab notebooks”: prototypes, experiments, and write-ups. Some will be polished; many are intentionally exploratory.
rag-my-pdf— minimal RAG over PDFs in Rust (Rig), used as a baseline to explore agent memory + grounding.zkvm-merkle-comparison— comparative zkVM work (when relevant to guarantees / verification).
(Links are in the pinned repos.)