1. Courses
  2. Machine Learning in Production
  3. Capstone

Capstone project for Machine Learning in Production

Run the demand forecast every night

Mangrove's supply chain team has approved the demand forecast after a successful pilot. Put it into production as if for all 320 stores: a nightly pipeline that extracts from the warehouse, validates the data, builds features, predicts 4 days ahead and publishes by 01:30; weekly retraining with tracked runs and a promotion gate; a fast endpoint for the Store app's ordering screen with a latency budget; monitoring with alerts that have owners and runbooks; a rollback that takes minutes; and a model card the supply chain team signs off. Use a public retail sales dataset (for example the M5 competition data) or a synthetic Mangrove dataset, scaled down to run on a laptop, and a simple model; the engineering around it is what is assessed.

What you hand in

  1. A code repository with the pipeline steps, a schedule (DAG or cron), a Dockerfile and tests, runnable end to end on sample data
  2. A serving endpoint with a latency budget and a fallback, and the batch publish step
  3. A monitoring plan or dashboard with metrics, thresholds, owners and runbooks, plus the retraining and rollback procedure
  4. A model card and a one-page ownership and approval summary

Check your work

Tick each item when your project meets it. Be strict: this is the rubric the reviewer uses.

  • Runs are reproducible from versioned inputs: run date parameter, pinned environment, data snapshot and tracked runs with lineage to the code
  • Steps are idempotent and publish last, so a retry or backfill never leaves a half-written run visible
  • Training and serving use the same feature code, data is validated before prediction, and late data is handled
  • The endpoint meets a stated latency budget and has a fallback; the batch run fits its time window
  • Monitoring covers pipeline health, input drift and forecast error, with festival-aware thresholds, owners and runbooks
  • Promotion goes through a gate, and rollback to the previous model takes minutes
  • The model card states intended use, out-of-scope uses, metrics by segment, limits and owners

0 of 7 checked

Model answer

The model answer opens once your capstone passes the review, so you can compare your work with an outline of a strong submission.

Hand in your capstone

Your links are checked automatically, then the work is scored against the rubric above, item by item, with comments. A pass counts toward the course certificate and shows on your profile.

Loading…