Add eight-module AI route, Kafka practice and current app screenshots

[siddath]

The engineering pack lacked a sequential AI curriculum and a Kafka consumer-lag lesson. This adds eight AI modules from LLM foundations through a production capstone, plus one Kafka lesson exclusively in Distributed systems. The expanded pack has 40 lessons; AI contains the eight modules followed by its two existing shared exercises. Original lesson records and the FDE route are unchanged.

The previous 31-lesson pack is retained byte-for-byte with an explicit preview control, so existing notebooks can be reopened without merging or resetting histories. New materials are original, fictional and intentionally unsolved. README features, route documentation and actual-application screenshots describe the resulting app and its limits.

Validation:

• 78 application tests pass, including ordered route IDs, prerequisites, legacy identity, pack isolation and tutor-context boundaries.
• Syntax and static build pass; the build includes 28 allowlisted assets and no backend/local data.
• Current-file and reachable-history publication audits pass; independent content and screenshot/privacy reviews pass.
• Browser checks cover all six filters, all nine new lesson entry points, AI/Kafka recognition and model gates, the previous/new notebook round trip, desktop and 375px layout, keyboard focus and manual reduced motion. No console warnings/errors or horizontal overflow observed.

The new exercise starters were syntax/fixture checked; their TODO tests intentionally fail until learners implement them. No live framework, provider, Kafka cluster or capstone deployment was run, and no learning achievement is claimed.