Private Team Training

Train your team on your architecture, not a demo tenant.

Small-cohort AI and data analytics training for 8–12 technical people. On-site or live online, taught by architects who build these systems for a living, using labs we build around your actual stack.

Request a quote

Tell us the shape of it and we’ll come back within one business day with a proposed outline and cost.

Private Team Training
Why private, not public

A public course teaches the tool. A private one teaches your system.

When the whole room works at the same company, the training stops being generic on day one, examples come from your data, questions get specific, and nobody has to translate afterward.

Your data, your constraints

We build the labs around your architecture and governance rules, so the examples aren't hypothetical and neither are the answers.

Questions people won't ask publicly

Nobody raises the messy legacy problem in a room full of strangers. In a private cohort, that's usually the most useful hour of the week.

The team levels up together

Shared vocabulary and shared standards across the whole group. Sending three people to a public course rarely changes how a team works.

Scheduled around your delivery

Full days, half days, or two afternoons a week for three weeks. We fit the sprint calendar rather than the other way round.

Content you keep

Lab guides, code, slides, and recordings stay with your team, usable for onboarding the next engineer who joins.

Cheaper past six people

Per-seat public pricing stops making sense around the sixth attendee. At 8–12, a private cohort is usually the less expensive option.

Cohort size

Why we cap it at twelve.

Twelve is the largest group where one instructor can still get to every screen during a lab. Past that, the people who are stuck stay stuck, and the session quietly turns into a lecture.

Below eight, you lose the discussion — the disagreements between a data engineer and an analyst about how something should work are often where the real learning happens.

Larger group? We’ll run parallel cohorts or bring a second instructor rather than stretch one past twelve.

delivery

On-site or live online. Same labs, same instructor.

Both formats run the same hands-on content in browser-based lab environments, nothing to install, nothing to configure on your machines.

on-site

We come to you

Best when the team is co-located and you want the focused, off-the-desk intensity of a room with the door closed.

Live online

Distributed teams, same room

Live and interactive, not a recording. Best for teams across time zones or when travel budget is the constraint.

Topic menu

What we teach.

Pick one, or tell us the outcome you need and we’ll assemble a custom track. Most engagements end up as a blend.

Most requested

Microsoft Fabric End-to-End

Lakehouse, pipelines, semantic models, and Direct Lake, the full path from raw source to a report someone trusts.

OneLake Medallion architecture Dataflows Gen2Direct Lake Capacity & Cost

3 days · Data engineers, analysts, BI developers

New

Building Production AI Agents

Go past the prototype: tool use, orchestration, evaluation harnesses, guardrails, and cost control.

Agent Framework MCP EvaluationGuardrails Monitoring

2 days · Developers, ML engineers, architects

AI

Microsoft Foundry in Practice

Model selection, prompt management, deployment, and monitoring inside a governed Azure environment.

Foundry Model Routing Content SafetyGuardrails Monitoring Observability

2 days · Developers, platform engineers

Data

RAG & Enterprise Search That Works

Why retrieval returns the wrong document, and the chunking, hybrid search, and re-ranking that fix it.

Vector Search Hybrid retrieval RerankingEval Sets Knowledge Graphs Semantic Search Semantic Caching

2 days · Developers, data engineers

Data

Databricks & the Lakehouse

Delta Lake, Unity Catalog, and production Spark patterns, including the ones that keep your bill sane.

Delta Lake Unity Catalog Spark tuning Workflows Genie

3 days · Data engineers

Analytics

Power BI & Semantic Modeling

DAX that performs, model design that scales, and governance so two teams stop publishing two versions of revenue.

DAX Star Schema Row-level security Deployment pipelines

2–3 days · Analysts, BI developers

Engineering

MLOps & Model Lifecycle

Versioning, CI/CD for models, drift monitoring, and rollback, treating models like software you have to operate.

MLflow CI/CD Drift detection Feature stores

2 days · ML engineers, DevOps

For leaders

AI Literacy for Technical Leaders

What to fund, what to ignore, how to read a vendor claim, and how to evaluate an AI proposal from your own team.

Capability mapping Build vs buy Risk & governance Cost models

Half day · Directors, VPs, architects

sampler agenda

What three days actually looks like.

Microsoft Fabric End-to-End, delivered to a mixed cohort of data engineers and analysts. Yours will differ, we rescope after the discovery call, but the shape holds.

Microsoft Fabric End-to-End

3 days · 8–12 attendees · On-site or live online
Day 1

Foundations & ingestion

09:00

Foundations & ingestion

OneLake, workspaces, capacities, and how the pieces relate to what you already run.

11:00

Lakehouse vs warehouse Lab

Build both, load the same data, and compare what each is good at.

13:30

Ingestion patterns Lab

Pipelines, Dataflows Gen2, and shortcuts, with the tradeoffs that decide which to use.

15:00

Medallion architecture in practice Lab

Bronze to silver, incremental loads, and handling late-arriving data.

16:15

Open floor

Your environment, your questions. Usually the most valuable half hour of day one.

Day 2

Transformation & modeling

09:00

Notebooks and Spark in FabricLab

PySpark transformations, when notebooks beat pipelines, and cost implications..

11:00

Gold layer & semantic models Lab

Star schema design, relationships, and the modeling decisions you can’t undo cheaply.

13:30

DAX that performs Lab

Measures, context transition, and diagnosing a slow visual with Performance Analyzer..

15:30

Direct Lake

How it actually works, when it falls back to DirectQuery, and how to keep it from doing so.

Day 3

Production, governance & AI

09:00

Deployment pipelines & source control Lab

Dev/test/prod in Fabric, Git integration, and what still has to be done by hand.

11:00

Governance & security Lab

Workspace roles, row-level security, sensitivity labels, and lineage.

13:30

Capacity, cost & performance

Reading the capacity metrics app, finding what’s burning CUs, and fixing it.

15:00

AI on your data Lab

Copilot, AI functions, and grounding a model on Fabric data, with the governance caveats.

16:15

Architecture working session

We whiteboard your real platform and leave you with a written set of recommendations.

How it works

From request to delivery in four steps.

01

Discovery call

Thirty minutes on your team’s current level, your stack, and what they should be able to do afterward.

02

Proposed outline

A written agenda, lab plan, and fixed cost, usually within one business day of the call.

03

We build the labs

Environments provisioned and content tailored to your architecture before anyone walks in.

04

Delivery & follow-up

Training runs, materials stay with your team, and we check in a month later on what stuck.

★★★★★

"This is THE best course I have taken so far. The instructor really knows the tool and is the most interesting, passionate instructor I have heard to date."

Nancy McGuire · Course attendee

questions

Before you ask for a quote.

What does it cost?

It depends on topic, length, format, and how much lab customization you want. Send the form and you’ll get a fixed price with the proposed outline, no hourly surprises. For most teams of 8–12, a private cohort works out cheaper than sending the same people to public courses.

Common, and workable. We send a short pre-assessment before the engagement and adjust pacing and lab difficulty from the results. Paired lab work also helps, stronger attendees consolidate their own understanding by explaining it.

Not necessarily. We provide sandboxed lab environments by default, so nothing touches your systems. If you’d rather train in your own tenant with your own data, we can do that instead, it makes the training more directly applicable but takes more setup on your side.

Yes, and for online delivery we often recommend it. Four half-days across two weeks usually retains better than two full consecutive days, and it doesn’t take the whole team out of delivery at once.

A laptop and a browser. Labs run in hosted environments, no installs, no admin rights, no VM downloads eating the first ninety minutes.

Yes, as parallel cohorts or with a second instructor. We don’t run a single instructor past twelve because lab support is what makes the format work.

Your team keeps the materials, lab guides, and recordings. We follow up about a month later to see what’s landed and what’s stalled. Teams that want ongoing support usually move to a fractional architect arrangement.

Tell us what your team needs to learn.

Send the shape of it, size, topic, format, timeframe, and you’ll have a proposed outline and a fixed price within one business day.

Scroll to Top