Skip to main content

GeekInDataBlog Explore Search Tools Author

Recent Posts

2026

POC public vs MVP private: networking as architecture for agent apps
Env files are not environments: isolate with Azure CLI discipline
Rename the estate once: workspace naming is architecture
Feature toggles beat hard-coded agent IDs
Language gates belong in CI, not in code review folklore
Genie readiness is a modeling problem, not a prompt problem
Type each column once, as early as you dare
Signing in is not authorization: app access is not agent access
Canvas apps at platform scale: push work to Dataverse
Legacy to lakehouse: semantic parity is the real migration
AI import UX is governance: default-deny uncertain rows
SCD2 is a product question: point-in-time deployed state with MLVs
Row counts lie: validate MLVs column by column
Fabric CI/CD blast radius: when no changes still deploys
Most NEE warnings are noise: measure real MLV wall-clock
Semantic refresh architecture: notebooks, pipelines, and Direct Lake
One Fabric lakehouse with schemas vs many lakehouses
Knowledge RBAC for enterprise agents: same gates for chat, browse, and citations
Consolidate Fabric monitoring into one operable skill
MLV migration is a program: bronze to silver to gold with gates
Bulk reload is an orchestration problem, not a bigger VM wish
Pilot the use case, not the agent catalog
If you cannot explain the tool calls, you do not have an agent product
Agent sessions vs the shell app: what should survive refresh
Foundry webapps deploy as identity and egress, not as Node hostnames
Refresh drift is a contract break, not a flaky pipeline
The AI shell proxy is a security boundary
PII redaction belongs in the notebook config, not in tribal knowledge
Pivot once in gold: measurable controls as a flat fact
Shortcuts-only bronze: stop copying what OneLake already holds
IoT events are facts: model them like service, not like logs
Matching relationship types is cheaper than debugging blank visuals
When warehouse views time out, materialize the trunk not the leaves
Variable Library is your environment map, not a sticky note
When Fabric regressions force literal OneLake URLs, document the workaround
Bridge keys before dashboards: remodel the device fact
Lineage explorers beat tribal notebooks
Gold without a semantic contract is just another table
Keep Fabric thin: Airflow orchestrates, Fabric does the work
Pose estimation is a systems problem, not a model bake-off
Classification apps fail on operations, not on the label list
Can Databricks Genie Build My Deck and Deploy My App?

Authors

Sai Prudhvi Neelakantam
42
Senior Consultant, Data Engineering & AI at Evidi

Site

Blog
Tools
Search
RSS
Author
LinkedIn newsletter

Tools

Archive
Answers
Post graph
Glossary
Pattern lab
Decision wizard
Compare
Developers / API

Copyright © 2026 GeekInData.