Contact
A conversation, when the problem is real.
Architecture, platforms, and where AI belongs. The same kind of problem, from firmware to cyber risk.
Based in Pune, India
The conversation
- 01An agent's trust boundary: which entities it may read, and which actions it may take
- 02An agentic step a product can commit, scoped to one entity, with an audit and a rollback
- 03RAG as a contract: governed sources, a freshness bound, and a citation back to the record
- 04LLM context cut to the decision, and a hard stop when the evidence runs out
- 05Voice held to the same entity a person can open and check
- 06Resolve the entity and retrieve the evidence before the model is allowed to generate
- 07Tool calling inside a blast radius the platform can audit and revoke
- 08A model may propose a score. The platform is what commits it.
- 09Risk on a business profile, assessed as two layers: the network, and the application
- 10That assessment as a platform, built in four to five days with Claude on the implementation and the scoring contract held in the architecture
- 11Company profiles at scale, every risk signal traced to its source
- 12Redshift held as separate contracts: landing, the model, and the mart
- 13One metadata definition that runs Spark, Python, and dbt
- 14Kinesis and Firehose into S3, live traffic and replay on the same field-level rules
- 15Redshift on the curated prefix as the query surface, Lake Formation where a subject must be deleted
- 16A regional copy that already carries its residency, its PII class, and its latency budget
- 17PII under data-protection law, enforced on the path: purpose, access, and deletion of every copy
- 18Cyber risk as evidence a decision can defend
- 19One entity question across firmware, chips, mobility, supply chain, banking, and healthcare
- 20A master record that products, marts, and agents all resolve to