Independent Researcher · Systems Engineer
I research and build systems for agentic AI assurance, explicit state transitions, runtime enforcement, deterministic execution, tamper-evident evidence, recovery, and verifiable infrastructure.
Published research
Develops a bounded execution-assurance framework for tool-using AI agents by coupling observation-bounded certification, typed pre-effect enforcement, claim-dependent evidence semantics, and a source-audited reference implementation.
Specifies a machine-verifiable execution contract for tool-using AI agents using externally observable runtime evidence rather than hidden chain-of-thought.
Research focus
A capable model does not automatically produce a reliable autonomous system. The surrounding runtime still has to decide whether an action is authorized, whether required invariants hold, whether execution is safe, whether resulting state can be verified, and how the system should recover after partial failure.
My work centers on the boundary between model intelligence and execution authority.
Selected engineering
Policy-enforced control plane for AI-agent tool execution and runtime evidence.
Deterministic transaction, operations, optimization, and evidence platform.
Experimental distributed SQL engine with explicit durability and recovery semantics.
Tamper-evident append-only evidence infrastructure with offline verification.
Proof-state-oriented symbolic reasoning and verifier-guided search.
Durable, provider-neutral coding-agent runtime focused on replay and safe resume.
Research identity
Independent Researcher & Systems Engineer