DVPApplied Creative AI Engineering

Search this course

Search lesson and reference material.

↑ ↓ to choose · Enter to open · Esc to close

Six-month course map

Applied Creative AI Engineering

A practical sequence for building inspectable creative-media systems: follow code, read frames, test detectors, and publish CineQC evidence.

Studio signal path
codeframesevidence
Film frames flow through arrays and measurements into validated detector evidence.
12 milestones82 guided hoursRelease 3: all four phases available
Local learning backup

Download a local copy of your progress, notes, and mastery state, or restore a backup from this course.

Notes

Your local study notes mount here when JavaScript is available.

Protected AI Tutor

Teaching modes mount here when JavaScript is available. The readable course remains available without them.

PHASE 1 / 04

Engineering independence

Read, build, debug, and verify code with confidence.

Available

M01 · Code as a story

Trace inputs, transformations, and outputs before you run anything.

6 hours · current: not-started

  • not-started
  • learning
  • practicing
  • evidence-submitted
  • proven
Available

M02 · Developer workbench

Build reproducible command-line habits for media tooling.

7 hours · current: not-started

  • not-started
  • learning
  • practicing
  • evidence-submitted
  • proven
PHASE 2 / 04

Visual computing

Turn pixels, frames, and media pipelines into inspectable systems.

Available

M05 · Image and video pipelines

Inspect and transform media without losing temporal context.

7 hours · current: not-started

  • not-started
  • learning
  • practicing
  • evidence-submitted
  • proven
Available

M06 · Rebuild the CineQC baseline

Measure artifacts before asking a model to learn them.

7 hours · current: not-started

  • not-started
  • learning
  • practicing
  • evidence-submitted
  • proven
PHASE 3 / 04

Learned detection

Build reproducible temporal-artifact models and evaluate them honestly.

Available

M08 · PyTorch and transfer learning

Train a small baseline you can resume and reproduce.

7 hours · current: not-started

  • not-started
  • learning
  • practicing
  • evidence-submitted
  • proven
PHASE 4 / 04

Product and proof

Integrate a detector into CineQC and publish defensible evidence.

Available

M10 · One semantic detector

Choose a defensible semantic target after the temporal benchmark.

7 hours · current: not-started

  • not-started
  • learning
  • practicing
  • evidence-submitted
  • proven
Practice and reference
Practice

Browser Python scratch lab

Run bounded Python, NumPy, and Pillow exercises, inspect output, and hand results to Tutor.

Evidence

Artifact Lab measurements

Five calibrated artifact metrics with published raw signals, thresholds, and false positives on controlled fixtures.

Evidence

Portfolio and case studies

Six competencies with committed evidence, reproduction commands, stated limits, and two written case studies.

Reference

Course map

The bounded route from code reading to product proof.

Reference

Python patterns for media tools

A compact lookup for paths, JSON, CLIs, contracts, and tests.

Reference

Safe terminal and Git field guide

Inspect first, preserve work, and make every change reviewable.

Reference

Shapes, color, and time field card

The compact frame and timeline checks that prevent silent media bugs.

Reference

Visual failure clinic

A symptom-first route for isolating image, video, and metric failures.

Reference

Evaluation metrics field guide

Counts, curves, calibration, and reproducible benchmark evidence.

Reference

PyTorch lifecycle field guide

The compact training, evaluation, checkpoint, and device loop.

Reference

Production integration checklist

Contracts, privacy, fallbacks, review routing, and release evidence.

Reference

Portfolio defense field guide

Turn inspectable artifacts into concise, honest competency proof.