Week 0 baseline (Sat 26 – Sun 27 Sep 2026)

Purpose: measure where you really are before you start, so the first weeks skip what you already know and target what you don't. Score everything with the unified rubric, and write the results in docs/log/baseline.md.

1. System design mock (60 min)

Prompt: "Design a URL shortener that handles 100 M new URLs/month with analytics". Use a timer: 5 min to clarify, 10 min to estimate, 20 min for high-level design, 20 min for deep dives, 5 min to wrap up. Record yourself if you can. Then compare against the case study.

2. Coding (3 × 30 min, Python, no IDE help)

• Longest Substring Without Repeating Characters
• Top K Frequent Elements
• Course Schedule

Note the time to a working solution, bugs found, and whether you explained the complexity unprompted.

3. AI engineering self-assessment (30 min)

Rate yourself 0–3 on each item (0 = never touched, 1 = read about it, 2 = built a toy, 3 = shipped to production):

Area	0–3
Calling LLM APIs with structured outputs (JSON schema / Pydantic)
Prompt caching, streaming, token/cost accounting
Embeddings, chunking, vector search, hybrid search, reranking
Building an eval set, error analysis, LLM-as-judge alignment
LangGraph (state, checkpoints, HITL)
Pydantic AI / OpenAI Agents SDK / Claude Agent SDK
MCP server/client, OAuth for tools
Tracing LLM apps (Langfuse / Phoenix / OTel GenAI)
Prompt injection defenses, OWASP LLM/Agentic Top 10
Serving open models (Ollama, vLLM), quantization
Fine-tuning (LoRA/QLoRA, DPO)
Spring AI

4. Environment setup (60–90 min)

• uv installed; uv python install 3.14 3.14t
• Docker Desktop / OrbStack running; docker compose works
• Ollama installed; pull a small model (e.g. a 3–8 B instruct model) and an embedding model
• API keys in a password manager and a local .env (never committed): one frontier-model provider + Azure subscription access
• JDK 25 (e.g. via SDKMAN) + Maven/Gradle
• Create the capstone repo agentic-ops-copilot (empty, with README)
• Clone this repo; uv sync && uv run mkdocs serve works

5. Communication baseline (about 3 h across the weekend)

Complete Week 0 drills: a 40-item grammar diagnostic, a 60-word vocabulary check, a recorded 2-minute talk, a 150-word status email and a self-assessment. Save the raw results in docs/log/comm-baseline.md; you'll repeat the same tests in week 24 and compare.

6. Write your skip list (30 min)

In docs/log/baseline.md, list:

• Skip / skim: topics you'd rate 3 with evidence (for each one, answer its L3 questions to confirm)
• Focus: everything rated 0–1 in AI, plus any DSA pattern that took more than 30 min
• Goals in your own words, and what "success" looks like on 14 Mar 2027