30-Day Prep Plan¶
Last reviewed: October 2026
A practical, four-week plan to go from "I know the topics" to "I can design, build, and defend production AI systems in an interview." Every day maps to specific OfferReady pages so you're never guessing what to study.
How to run it
~1–2 focused hours/day. Each day = learn → practice out loud → self-quiz. Reading isn't rehearsing; say answers aloud and time yourself (~2 min). Track yourself in Practice mode and the Progress Dashboard (both now open in the OfferReady app). Adjust weeks to your target role (see Level Comparison).
Daily rhythm: 1) read/skim the concept page · 2) do the linked drill/Q&A out loud · 3) run the page's self-quiz · 4) note weak spots to revisit on the review day.
Week 1 — LLMs, RAG & AI fundamentals¶
Goal: fluent on the building blocks and retrieval.
| Day | Focus | Pages |
|---|---|---|
| 1 | LLM fundamentals (tokens, sampling, context, tool calling) | LLM Fundamentals |
| 2 | Prompt + context engineering | Prompt · Context |
| 3 | RAG end to end | RAG · Retrieval Tuning |
| 4 | Embeddings + vector/graph DB | Embeddings · Vector DB |
| 5 | AI Engineer Q&A (RAG/eval/cost) | AI Engineer Q&A |
| 6 | Practice: RAG scenario + a design | Scenario Drills · Requirements → Production |
| 7 | Review week 1 weak spots + Why-chains: RAG/model | Cheat sheet: LLM/RAG |
Week 2 — Agents, MCP & Security¶
Goal: design safe, reliable agentic systems and defend them.
| Day | Focus | Pages |
|---|---|---|
| 8 | Agent principles + patterns (deep) | Agent Principles |
| 9 | LangChain / LangGraph | LangGraph · LC/LG Q&A |
| 10 | MCP + tools (incl. the stateless 2026-07-28 spec, OAuth, tool poisoning) | MCP · MCP Q&A |
| 11 | AI Security (injection, tool poisoning, threat modeling) | AI Security |
| 12 | Agents Q&A (context engineering, agent evals) + a multi-agent design | Agents Q&A · Case Studies |
| 13 | Practice: agent + security scenarios | Scenario Drills · Why-chains: agents/security |
| 14 | Review + Hackathon build (agent/MCP challenge) | Cheat sheet: Agents/MCP/Security |
Week 3 — Data & Cloud platform¶
Goal: depth on the data/cloud stack around AI.
| Day | Focus | Pages |
|---|---|---|
| 15 | SQL (windows, tuning, modeling) | SQL Q&A · Live-Coding Drills |
| 16 | Data engineering / pipelines | Data Engineering Q&A |
| 17 | Snowflake + Cortex | Snowflake · Snowflake Q&A |
| 18 | Databricks / Spark / Delta | Databricks Q&A |
| 19 | dbt + AWS | dbt Q&A · AWS Q&A |
| 20 | Python for AI + live-coding drills | Python Q&A · Live-Coding Drills |
| 21 | Review + Why-chains: data platform | Cheat sheet: Snowflake/Databricks/SQL/Python |
Week 4 — System design, ops & production¶
Goal: put it together at senior/principal depth and rehearse full loops.
| Day | Focus | Pages |
|---|---|---|
| 22 | LLMOps + Production CI/CD (Kiro/Jenkins) | LLMOps · DevOps for AI |
| 23 | Observability + Reliability | Observability · Reliability |
| 24 | Kubernetes + cost/performance | Kubernetes · Cost |
| 25 | Production incidents | Production Incident Interviews |
| 26 | Full system design | Requirements → Production · Case Studies |
| 27 | Behavioral / STAR story bank | Behavioral / STAR |
| 28 | Full mock loop | Mock Interview (OfferReady app) |
| 29 | Level-tune your answers (Senior→Staff→Principal/FDE) | Level Comparison |
| 30 | Final review: cheat sheets + retake weak self-quizzes | all self-quizzes |
Track adjustments¶
Weight the weeks toward your target role (see Level Comparison):
- AI / GenAI Engineer: full plan as written.
- Data / Analytics / Snowflake / Databricks Engineer: double Week 3; keep Weeks 1–2 lighter (RAG + agents awareness).
- Forward-Deployed Engineer: add the FDE pages and FDE Live-Coding across Weeks 2 + 4; emphasize Requirements → Production and communication.
- Staff / Principal: spend extra time on Week 4 (system design, cost, governance) and the Why-chains.
The week-before checklist¶
- Can design a full system from Requirements → Production out loud
- Can defend each decision 5 "why"s deep (Why-chains)
- Can walk 3 production incidents end to end
- Have 6–8 STAR stories ready
- Skimmed the cheat sheets the night before