How much semantic modeling do your AI agents need?
A meeting room booking example shows when SQL is enough, when a property graph helps, and what would justify RDF or OWL.
Writing
Articles and notes
Hands-on articles and shorter notes. New writing is published here first, then on LinkedIn and Medium. Everything is in the RSS feed.
Series
Beyond Similarity
A series on ontology-grounded agents. Retrieval finds text that resembles a question. An enterprise needs the records that satisfy it, with evidence.
- Part 1Why AI agents need clear business definitionsSep 30, 2026
- Part 2Why AI audit agents need more than search resultsOct 1, 2026
- Part 3How much semantic modeling do your AI agents need?Oct 4, 2026
Why AI audit agents need more than search results
In a synthetic test, an AI agent returned 3 correct patients and quietly missed 9. Here is why that happened, and how to check whether a list of your own is complete.
Why AI agents need clear business definitions
A grocery substitution explains ontology, knowledge graphs, semantic and context layers, and how business and technology teams keep their definitions trustworthy.
How to configure OAuth2 authentication for an Apache Kafka cluster using Okta Medium
The OAUTHBEARER SASL mechanism with custom callback handlers that validate tokens against Okta, wired to ACLs, tested end to end. Companion repo: kafka-oauth2.
How to configure KSQL, a streaming SQL engine on Apache Kafka Medium
Streams versus tables, keys and timestamps, filters, joins and aggregations, without writing a Kafka Streams topology. Docker based.
Configuring Kafka Burrow for a SASL Kafka cluster Medium
Consumer-lag monitoring with LinkedIn's Burrow, including the fork that added SCRAM-SHA-256 and SHA-512 support, a compose setup and a dashboard.
Securing an Apache Kafka cluster with SSL, SASL and ACLs Medium
SCRAM-SHA-512 authentication, ACL authorization and SSL encryption, with the errors you hit along the way and how each one is fixed. Companion repo: SecureKafka.