Elie Hasbani

Work / BigBrain Knowledgebank

BigBrain Knowledgebank

Partial 2026

A provenance-tracked knowledge graph of the estate: what each project taught, where it was learned, and whether it is still believed.

The real graph (node types and edges; the layout is seeded) and one real retrieve, use and write-back loop from the run ledger. Only that task's own nodes are labelled.

What it does

Every accepted piece of AI work writes back what it learned: a fact, gotcha, decision or lesson, tied to its project and its evidence. Before the next task, the relevant nodes are retrieved and injected into the supervisor and workers within a token budget, and each worker reports which ones it actually used. Writes are transactions with before-and-after SHA-256 hashes and a byte-exact rollback, so a bad memory can be removed for good.

How it runs

  1. Accepted work writes back a fact, gotcha, decision or lesson
  2. Each write is a transaction with before-and-after hashes
  3. The next task retrieves the relevant nodes within a token budget
  4. Supervisor and workers receive them with the brief
  5. Each worker reports which nodes it actually used
  6. Wrong or outdated knowledge is superseded or rolled back

How it works

The graph holds 1,190 nodes and 1,250 edges: projects, modules, routes, decisions, gotchas, patterns, lessons and facts, each with its project, evidence type and trust level. Retrieval is measured, not assumed: on 13 held-out questions it found the right node in the top three 0.917 of the time, with no polluted results. Newer knowledge supersedes older knowledge rather than overwriting it, so the record of what was believed survives. The rollback was rebuilt after an audit found the earlier one existed only on paper: the documented delete route was missing, and a restart restored the node.

Skills shown

Data engineering
A provenance-tracked knowledge graph where newer knowledge supersedes older knowledge instead of overwriting it.
Security
Every write is a transaction with before-and-after SHA-256 hashes and a byte-exact rollback.
AI, agents & ML
Retrieves the relevant knowledge into model prompts within a token budget, and records which pieces each worker actually used.
Testing & QA
Retrieval quality is benchmarked on held-out questions rather than assumed.

Worth knowing

  • Retrieval is benchmarked on held-out questions: the right node is in the top three 0.917 of the time, with zero polluted results.
  • Workers report which injected nodes they applied, ignored or found stale. The recorded application rate is 0.348.
  • New knowledge supersedes old knowledge instead of overwriting it, so the history of what was believed stays inspectable.
  • Every write is a transaction with a durable backup and a byte-exact rollback.