Harness

Harness & Production AI

The machine around the model: typed tools, an explicit runtime loop, context, state and memory, permissions and human approval, verification, tracing, evals and failure injection.

0 published · 12 planned

The machine around the model

SystemsThe Harness: The Machine Around the ModelComing 16 Dec

LabHarness 01: Build a Typed Tool RegistryComing 19 Dec

SystemsInside the Agent Execution LoopComing 23 Dec

LabHarness 02: Build an Agent Runtime From ScratchComing 26 Dec

Context, state and memory

SystemsContext Is Infrastructure: Retrieval, Live State and Context BudgetsComing 30 Dec

LabHarness 03: Build a Context + RAG PipelineComing 2 Jan

SystemsMemory Is Not State: Designing Persistence for AI SystemsComing 6 Jan

LabHarness 04: Add Persistent State and MemoryComing 9 Jan

Control and evidence

SystemsWho Is Actually in Control? Permissions, Policy and Human ApprovalComing 13 Jan

LabHarness 05: Build Permissions + Human ApprovalComing 16 Jan

LabHarness 06: Verification and EvidenceComing 23 Jan

LabHarness 07: Production Evals, Tracing and Failure InjectionComing 30 Jan

Other libraries

Intelligence from Zero →Fine-Tuning Lab →AI Integration for Companies →IntelligenceOS Build Journal →