Bio
Brandon Callender is a founding engineer at typedef, where he builds AI-native infrastructure for data engineering agents. His work focuses on the data context layer agents need to reason beyond code and database access.
Session (1)
Additional Speaker Details
Using publicly available information we constructed an analysis to help you get a feel for this speaker before deciding to attend their session.
His personal GitHub project agent-context-workshop is a rigorously benchmarked comparison of grep-based vs. graph-based context retrieval for coding agents — pitting a graph agent against a grep agent on 30 verified questions about the qdrant codebase and measuring accuracy, determinism, and token efficiency. It reflects a technical, evidence-driven approach grounded in real experimentation, consistent with his work as an engineer at typedef, which builds a persistent 'data context layer' for AI agents.
Speaking style
GitHub
@bcallender- bcallender/agent-context-workshop★ 15A context layer for coding agents: builds a code knowledge graph from a real codebase and measures whether an agent actually uses it, vs. grep-based retrieval.
- bcallender/AllAboard★ 10Cities: Skylines II mod that forces transit vehicles to depart on a max dwell-time limit instead of waiting indefinitely for boarding.
- bcallender/raft★ 1Raft consensus protocol implementation (coursework, CMSC 23310).