This ASIMOV DevLabs session builds on the previous workshop, focusing on how structured knowledge graphs enable personal, context-aware intelligence systems that remain local, private, and user-controlled.

Guest lightning talk:
Rob Kunkle — a software architect, documentation engineer, and knowledge systems engineer — has built large-scale documentation systems at Google, Salesforce, and Meta, including the PyTorch v1 documentation.

Rob will demo his work on Repolex, a system that transforms code repositories into composable, queryable graphs — AST, LSP, dataflow, and control flow — accessible via SPARQL. The goal: give AI agents richer structural understanding of codebases than agentic search alone can provide, with graphs that can be downloaded directly from Git. Rob will walk through the engineering decisions behind the project and how it integrates with ASIMOV's knowledge architecture

Arto Bendiken will walk through how ASIMOV’s architecture supports multiple intelligence layers - perception, identity, memory, and reasoning - unified through verifiable knowledge models rather than centralized data extraction.

Participants will explore real-world use cases, extend graph structures, and examine how local-first intelligence enables new classes of applications without compromising provenance or agency.

This workshop is part of ASIMOV’s ongoing Bay Area DevLabs "Context Graphs & Personal Intelligence" series.

Who this is for

Developers and researchers working with:

Knowledge representation
Graph-based systems
AI infrastructure and tooling
Privacy-first or local AI

Hosted by Circuit Launch.

ASIMOV DevLabs 5: Context Graphs & Personal Intelligence