Databricks training your team uses on Monday

Lakehouse Fundamentals

For teams new to Databricks or migrating from a warehouse. Build a working lakehouse in a day.

• Delta Lake: ACID tables, time travel, schema evolution
• Medallion architecture done properly
• Clusters, compute types, and when serverless wins
• Databricks SQL and dashboards
• Orchestration with Jobs and Lakeflow

Unity Catalog & Governance

For platform teams who need governance that auditors accept and engineers tolerate.

• Metastore, catalogs, schemas: a layout that scales
• Grants, groups, and least-privilege patterns
• Row filters, column masks, and lineage
• Migrating off hive_metastore without downtime
• External locations, volumes, and Delta Sharing

Performance & Cost Tuning

For teams whose DBU bill grew faster than their data. For teams with a six-figure annual DBU bill, one avoided mistake can cover the workshop cost.

• Anatomy of a DBU bill: where the money actually goes
• Cluster policies, autoscaling, and spot strategies
• Photon and serverless: when they save, when they don't
• Reading the Spark UI: skew, spill, and shuffle
• Delta file layout: OPTIMIZE, liquid clustering, VACUUM

Gen AI on Databricks / Mosaic AI

For teams under pressure to ship AI features without shipping an incident. Includes AI security throughout.

• Mosaic AI Model Serving and foundation model APIs
• Vector Search and RAG pipelines on your documents
• Agents, tools, and the Mosaic AI Agent Framework
• Evaluation: judging quality before your users do
• AI Gateway, guardrails, and prompt-injection defense

MLOps with MLflow

For data science teams whose models work in notebooks and nowhere else.

• Experiment tracking that survives team turnover
• The model registry in Unity Catalog
• Deployment patterns: batch, streaming, Model Serving
• Monitoring, drift, and retraining triggers
• Feature engineering and reproducible pipelines

Something else?

Streaming with Structured Streaming, DLT/Lakeflow pipelines, Databricks on Azure specifics, Terraform for workspaces — if it's Databricks and I run it in production, I can teach it. Tell me the gap. Upskilling a single engineer rather than a team? 1:1 Databricks mentoring is usually the better fit — or hands-on Databricks consulting if you need it built with you rather than taught.

Answers from production

When someone asks "but what happens when the metastore migration hits a legacy mount?", they get an answer from a person who has done that migration — with the failure modes, not the marketing version.

Certified in what I teach

I hold the professional-level Databricks certifications in data engineering and machine learning, plus the Gen AI Engineer Associate. I've sat the exams your team will sit, recently enough to remember them.

Opinionated, with reasons

You won't hear "it depends" and a shrug. You'll hear what I'd do on your platform, why, and what I'd trade off — the judgment your team is actually missing, not another feature tour.