02SELECTED WORK / PUBLIC ARTIFACTS

Case studies without
invented outcomes.

Each case names the problem, what was built, the evidence available for inspection, and the boundary that evidence does not cross.

CASE / 01

Mnemes and semantic-memory

Problem: Agent memory must persist, retrieve relevant prior state, preserve provenance and time, and remain operator-owned.
Built: A Rust-first memory engine, MCP access layer, agent integration kits, and self-hosted Mnemes server with explicit ownership boundaries.
Evidence: Public repositories, crates, documentation, tests, and a dedicated product/proof surface.
Boundary: Public artifacts establish implementation scope, not customer adoption, universal correctness, or production fitness.

Inspect semantic-memory ↗

CASE / 02

Hermes integration path

Problem: A capable Python agent stack can become resource-heavy when persistent memory, services, and multiple agents run together.
Built: A RecursiveIntell-enhanced Hermes path integrating Rust services, memory tooling, MCP installation, and multi-agent orchestration.
Evidence: Public installer source and a public demonstration that Hermes creator Teknium highlighted on August 5, 2026.
Boundary: The interaction is not a partnership or testimonial, and the public demonstration is not an independently reproduced benchmark.

View the public interaction ↗

CASE / 03

ClaimLedger and agent-graph

Problem: Agent plans and claims lose value when evidence identity, promotion rules, execution history, and failure states remain implicit.
Built: Public Rust components for claim/evidence history, typed graph execution, receipt-bearing tool paths, and bounded orchestration.
Evidence: Published crates, public source, schemas, examples, and test surfaces that can be inspected directly.
Boundary: Receipts report observed execution. They do not prove factual truth, authorization, security, or task success.

Inspect claim-ledger ↗

CASE / 04

turbo-quant and compression research

Problem: Resource-constrained AI needs smaller representations without hiding formats, quality budgets, or exact fallback paths.
Built: Experimental Rust quantization and compression primitives with explicit codecs, benchmark surfaces, and raw/exact escape hatches.
Evidence: Published package records and public source for turbo-quant and related components.
Boundary: No comparative performance or memory-savings claim is published here without a reproducible, revision-bound benchmark.

Inspect turbo-quant ↗

THE PROOF LAW

Evidence has
a defined scope.

A repository can establish source exists. A package can establish publication. A test can establish its declared invariant for one revision and environment. A receipt can establish what a run observed. None inherits customer adoption, production fitness, security, factual truth, or business outcomes for free.

Explore the full portfolio →Apply the patterns through consulting →