Bradford Chang
Bradford Chang works in field engineering at Databricks and teaches developers to coordinate coding-agent workflows, including pull-request review, repair and validation. Previously at AWS, he has discussed both data infrastructure and the practical obstacles to using agents: switching between coding tools, managing token consumption and authenticating access. The AI Engineer Code 2026 roster lists him as a senior solutions architect at Databricks.
On this page
From lakehouse infrastructure to coding agents
Bradford’s earlier public posts addressed lakehouse technology and adoption. In 2022, he shared comparisons between Delta Lake and Iceberg, offered to help developers understand Delta, and welcomed Databricks’ availability through AWS Marketplace. These posts concerned evaluating storage formats and making a data platform easier to adopt; they do not establish authorship of either technology.
More recently, Bradford has used and promoted Omnigent, an open-source orchestration layer that brings multiple coding-agent harnesses into a shared interface. In June 2026, he described replacing manual switching between harnesses with one place to work, accessible from his phone. His interest also includes the cost of coordinating agents. When Omnigent introduced smart routing, he wanted to measure its effect on token consumption; he did not present savings as an established result. He has separately identified authentication as a significant difficulty for agents.
Connecting review to tested repairs
For November 2026, Bradford and Elise Gonzales are scheduled to lead an Omnigent workshop on pull-request review and repair. Their planned workflow takes review comments through repairs and tested fixes, with explicit attention to agent permissions and validation. This gives Bradford’s teaching a concrete application: coordinating agents across a piece of software work that must end with a checked change.