Learn AI by operating a mobile network.
Move from a network alarm to Python investigation, statistically defensible findings, and a trusted data pipeline. Every workbench uses the same fictional operator and versioned evidence.
Evening congestion is affecting the coastal cluster.
Your task is not merely to detect a red indicator. Confirm the quality of the data, prove the scale of the problem and submit an engineering recommendation.
Network traffic & capacity pressure
Last synthetic day, hourly mean across available cells
Operational response
Move one incident through three skills
Priority incidents
Ground-truth events inside the simulator
Learning progress
Three practical courses, one continuous capstone
Radio network explorer
Explore simulated site and cell health. Select a site to filter subsequent statistical investigations.
Site health matrix
24 sites / 72 cells · selected scenario's most recent interval
Site detail
Select a location
Selected site · 24-hour KPI profile
Actual arithmetic on generated telemetry; not fabricated chart output
Capacity decision simulator
Explore a human-reviewed response to the selected site. Hypothetical scenarios only: no configuration writes.
Incident investigation desk
Investigate a fault, acknowledge the alarm, document evidence and choose the correct root-cause hypothesis.
Open operational tickets
Simulated events, fixed ground truth, local ticket states
Incident details
Choose an incident
Python engineering studio
Use the real Python interpreter in the browser to inspect a sampled telecom CSV. Program output and competency tests are executed, not hardcoded.
Engineering tasks
Complete 4 real code challenges
csv_text. Runs use a locked sample to make grading reproducible.Task 1: Count records
Load your sample and get started
Load Python runtime, then run your code.
Network data inspector
Reproducible snapshot from the simulator
Statistics & data literacy
Move beyond KPI averages: inspect distribution tails, relationships, confidence bounds and paired operational comparisons.
Hourly KPI profile
Means aggregated by hour for the selected population
Distribution & decision evidence
Not every high average indicates a fault
Paired peak / off-peak investigation
Each cell-day contributes one paired mean; normal approximation is illustrative
Four applied competency checks
Use calculated evidence to complete the statistical course
Telecom ETL & data quality studio
Profile raw performance exports, configure quality rules, remove duplicates, quarantine invalid observations and publish versioned, traceable KPI data.
Build the quality pipeline
These switches change real JavaScript transformations
Schema and quality gates
Engineering rules, not cosmetic badges
Data preview
Raw telemetry · first 8 records
Lineage & pipeline audit
Track what changed, why, and which dataset was published
Course 03 checkpoints
One operator, three linked courses
Learn the complete investigative cycle. Mission progress is stored in this browser; each course contributes evidence to the final report.
Real-world operating workflow
The same operational event moves across three disciplines
Investigation & capstone report
Combine validated KPI calculations, incident observations, pipeline versions, and completed training competencies into a traceable engineering report.
Technical report preview
Calculated when you open this page
Engineer commentary
Separate observations from recommendations
Data sources & integration readiness
Learn how a real operator would connect OSS/PM, alarm and asset systems. This standalone HTML has no authenticated live-network integration.
Operator ingestion paths
Proposed backend adapters, not active integrations
Bring your own KPI export
Use a CSV with the defined canonical schema in the ETL sandbox
Required: timestamp_utc, site_id, cell_id, prb_dl_util_pct and dl_throughput_mbps. Extra known fields are optional. Max 6 MB. Uploaded CSV stays in your browser.
Production architecture — implementation stages
Network Planning & Field Lab
Analyze the same virtual operator across all 30 sessions. Move customers, plan sites, upgrade capacity and inspect coverage changes. Publish modeled KPI impacts to the NOC simulation.
Network engineering comparison
Before and after the selected intervention, evaluated against the same synthetic subscriber distribution
Engineering work packages — 30 linked sessions
Each session uses this shared network case, measurements and project evidence
30 sessions. 300 topics. 30 operator missions.
Three full courses inside one 4G/5G operator simulation. Every session includes ten practical topics, a role-specific engineering case, downloadable evidence, automatic calculation checks, and a documented recommendation.
Telecom engineering investigation
Real calculations · Practical evidence · No simulated grading outcomes
10 learning topics
Write a field note to record your practice
Network evidence workbench
Current synthetic or imported operator data, filtered to this session's scope
Complex operator assignment
Quantitative verification + evidence review