Interrogate the estate
Use SQL to establish grain, point-in-time availability, and reconciled evidence before reaching for a model.
SIMULATION WORLD / MERIDIAN LIVING SYSTEMSFROZEN ENTERPRISE ESTATE / REV 2026.09
ASSIGNMENTS09foundation to practicum
ENTERPRISE SCHEMAS1696 documented tables
FULL ESTATE16.5Msynthetic source rows
RUNTIMELOCALSQL + Python in browser
01 / THE SIMULATION WORLD
YOUR EMPLOYER, FOR THE DURATION OF THE WORK
Welcome to Meridian. Your inbox is already full.
The Analyst is the training environment. Meridian Living Systems is the fictional company inside it: a multi-region operator selling connected-home products and subscriptions, dispatching field service, running warehouses and commerce, and supporting customers after the sale.
Meridian has grown through acquisitions, platform migrations, new models, and operational change. Its data reflects that history. Across nine assignments, you move between teams and moments in the same company. The systems, customers, devices, branches, definitions, and consequences belong to one connected estate—not nine unrelated classroom datasets.
02 / OPERATING MODEL
A workbench, not a walkthrough.
Learners receive a role, an imperfect request, a governed data neighborhood, and a deadline. The system supplies tools and evidence—not a hidden preferred conclusion.
Analyze in Python
Profile, visualize, test, backtest, simulate, or model with the package profile appropriate to the assignment.
Leave an audit trail
Export the exact SQL, Python, working notes, final brief, evidence, captured outputs, versions, and hashes in one portable submission.
03 / OPTIONAL AI ASSISTANCE
Two useful tools. One non-negotiable standard.
AI assistance is optional. The workbench remains fully usable without it, and generated advice is never treated as evidence.
A bounded second opinion
Across every assignment and Priority Brief, a small local model can explain a visible error, compare a highlighted recommendation with the request, or pressure-test a highlighted technical claim. It is not an open chat, grader, or answer key—and no account is required.
Take reviewed context elsewhere
Every assignment and Priority Brief can produce a structured Markdown packet for an external AI tool. You choose the question and included material. The packet is created locally and leaves the workbench only when you deliberately share it.
Advice must survive verification
Use AI to explain, question, and surface checks—not to certify a result or decide what the evidence means. Run the analysis, inspect the sources, verify important guidance, and make a handoff you can defend.
04 / ASSIGNMENT SEQUENCE
Built to progress.
The sequence moves from metric reconciliation to production-grade model risk. Complexity rises, but judgment begins in assignment one.
CONNECTED RELIABILITY / INCOMING
The Silent Sensors
Device readings look calmer as telemetry coverage falls during a storm and firmware change. Decide whether the apparent recovery can narrow containment.
- YOUR ROLE
- Reliability Duty Analyst
- WORKING TIME
- 75–105 minutes
- DATA NEIGHBORHOOD
- 6 existing tables / 110,362 mounted rows
- DIFFICULTY
- Intermediate
FOR INDEPENDENT LEARNERS
Run a serious course of study—without pretending you have an instructor.
The self-guided manual supplies readiness gates, honest workload routes, a repeatable assignment cadence, stuck protocols, and a spoiler-controlled way to review your own work.
CAREER READINESS / STUDENTS + CAREER CENTERS
Turn completed work into evidence you can defend.
The Career Readiness Evidence Guide connects assignments to carefully bounded NACE competency practice, portfolio review, truthful résumé and career-profile language, interview preparation, and ready-to-use career-center formats.
FOR INSTRUCTORS AND FACILITATORS
Planning guidance, workload bands, assessment boundaries, and local review tools sit in a deliberately separate instructor desk so the learner route stays focused on the work.
MONDAY / 08:05 / CUSTOMER CARE