AI Governance Platform
An open-source platform for governing, evaluating, replaying, and operating AI systems-with reproducibility and operational control treated as first-class concerns.
Project notesMELBOURNE / AI Architecture
Bhanuj
Bawdhane.
AI Systems, Data & Architecture
I design and build enterprise AI systems, with particular interest in governance, evaluation, reproducibility, data architecture, and operational reliability. Building software since 2011.
01 / FEATURED WORK
Engineering Projects
02 / WRITING
Long-form technical work.
07 Aug 2026
→Lessons From Building AI Governance Platform
Building a governance and operational-control platform for AI systems reinforces a simple lesson: evidence, policy, and replay must be designed into the system path.
06 Aug 2026
→Iceberg as an Immutable Operational Ledger
Apache Iceberg is most useful when its snapshot model is treated as an operational record of data behaviour—not merely as a faster table format.
05 Aug 2026
→Lakehouse Architecture Beyond the Hype
A lakehouse is not a storage format or a vendor category. It is an operating model for data, with explicit ownership, contracts, lifecycle, and query responsibility.
04 / ENGINEERING PRINCIPLES
Clarity is a feature of the system.
01
Operational stability over elegance
Design for diagnosis, recovery, and change-not just a clean diagram.
02
No hidden contracts
Make APIs, schemas, events, and behaviour explicit and versioned.
03
Failure isolation
Independent components should fail independently wherever practical.
04
Reproducibility
An outcome should retain the model, policy, data, and configuration that shaped it.