Previous lesson. Capstone: ML Engineer: Real-Time Risk Scoring System (GDPR-Safe) · Next lesson. Project 1: Image Classification (CIFAR-10)

Capstone Blueprints (public portfolio practice)

These capstones are portfolio-oriented project blueprints for public GitHub practice (ML Engineer / LLM Engineer / Data & Analytics role paths). They are practice targets only. They do not guarantee a hire. They are written to be safe for a public GitHub repo:

Use public datasets (download yourself) or synthetic data
Do not commit datasets, model artifacts, logs, or credentials (this repo’s .gitignore already enforces that)
Avoid any personally identifiable information (PII) or private company documents

Note: This is educational content, not legal advice. See DISCLAIMER.md for data/privacy context.

Capstones

1) ML Engineer Capstone: Real-Time Risk Scoring System

Blueprint: capstone-ml-engineer.md
Focus: end-to-end pipeline, API serving, monitoring, retraining, cost/reliability thinking

2) LLM Engineer Capstone: RAG Knowledge Assistant (with evaluation + guardrails)

Blueprint: capstone-llm-rag-engineer.md
Focus: ingestion, embeddings, retrieval, generation, eval harness, prompt-injection defenses

3) Data/Analytics Capstone: SQL, Metrics, Dashboard, ML

Blueprint: capstone-data-analytics-sql-ml.md
Focus: analytics engineering, SQL case studies, business metrics, and a productionized ML model

No PII: no names, emails, addresses, phone numbers, IPs, exact GPS, device identifiers
No secrets: .env stays local; never commit API keys/tokens
No datasets committed: only provide download links or synthetic generators
No raw logs: logs can contain identifiers; keep them local
Document tradeoffs: explain privacy decisions and data minimization in your README

All lessons in this module