The 8-Week Plan

Python, fast — through PHP eyes

Goal: read and write idiomatic Python without thinking in PHP.

• Syntax shift: indentation, no $, f-strings, truthiness
• Lists, dicts, comprehensions; functions & *args/**kwargs
• Type hints, virtualenv + pip, modules & imports
• Classes & dataclasses (you'll need them for Pydantic next week)

Primary: From PHP to Python · Official tutorial → see Lesson 0001

FastAPI — your first real API · Lesson 02 →

Goal: build & document a CRUD REST API from scratch.

• Routes, path/query params, request/response models
• Pydantic for validation (this is the FastAPI superpower)
• async/await basics; status codes & error handling
• Auto-docs (Swagger), project structure, dependency injection

Primary ⭐: FastAPI official Learn

Interview hook: “REST API design, validation, why async.”

Postgres + SQL — make it persist · Lesson 03 →

Goal: wire the API to a real database with auth.

• SQL fundamentals: tables, joins, indexes (you'll be quizzed on these)
• SQLAlchemy 2.0 ORM + Alembic migrations
• Connect FastAPI → Postgres; sessions & queries
• JWT auth: register / login / protected routes

Primary ⭐: SQLAlchemy 2.0 · Postgres tutorial

Interview hook: “SQL joins, indexing, ORM vs raw SQL, auth flow.”

Modern React — refresh the right way · Lesson 04 →

Goal: replace your jQuery instincts with hooks + components.

• Function components, JSX, props, the mental model shift from jQuery
• Core hooks: useState, useEffect; lifting state
• Fetching data from your FastAPI backend (fetch/axios)
• Lists, forms, conditional rendering, component composition

Primary ⭐: react.dev

Interview hook: “State management, useEffect, components vs the old DOM way.”

Next.js App Router — the full-stack frontend · Lesson 05 →

Goal: a production-shaped frontend for DocChat.

• App Router: file-based routing, layouts, server vs client components
• Data fetching & server actions; loading/streaming
• Connecting to your FastAPI backend; env vars
• Tailwind for fast, decent-looking UI

Primary ⭐: Next.js Learn (Dashboard)

Interview hook: “Server components, SSR vs CSR, when/why Next.”

RAG — the feature that gets you hired differentiator · Lesson 06 →

Goal: answer questions over uploaded docs using AI.

• The RAG loop: chunk → embed → store → retrieve → generate
• pgvector: vectors in Postgres, cosine distance, indexes
• Embeddings + an LLM call (Claude / OpenAI) for the answer
• Wiring retrieval results into the prompt (context injection)

Primary ⭐: RAG on Postgres (2026) · pgvector

Interview hook: “Explain RAG end-to-end, why vector search, chunking trade-offs.” Most candidates can't. You will.

A bit of DevOps — put it online · Lesson 07 →

Goal: a public URL you can paste into an application.

• Git/GitHub workflow & a clean README (recruiters read these)
• Docker: containerise the FastAPI backend
• Env vars & secrets; managed Postgres (Neon/Supabase)
• Deploy: Next.js → Vercel, backend → a container host; basic CI

Primary: Docker · Vercel

Interview hook: “How you deploy, env/secrets, CI basics.”

Polish, prep & apply · Lesson 08 →

Goal: stop learning, start interviewing.

• Polish the capstone: README with screenshots, architecture diagram
• Write your "project story" — be able to talk through every layer in 3 min
• Drill common full-stack interview Qs across the whole stack
• Update LinkedIn/CV; start applying via referrals

Interview hook: the whole project is the answer to “tell me about something you built.”

TypeScript — the type-safe stack · Module 9 →

Goal: write the React + TypeScript almost every UAE spec asks for.

• Types, inference, interfaces vs type aliases, unions & narrowing
• Generics and the essential utility types (Partial, Pick, Omit, Record)
• Typed props, hooks, events; zod + z.infer; typed Server Actions

Interview hook: “any vs unknown, generics, how you type an API response.”

Testing — ship with confidence · Module 10 →

Goal: pytest the backend, Vitest/Playwright the frontend.

• pytest: fixtures, parametrize, mocking the boundary, async, coverage
• FastAPI tests with TestClient/AsyncClient and dependency overrides
• React Testing Library, MSW, and Playwright end-to-end

Interview hook: “what do you test, what do you mock, the testing pyramid.”

AI Agents & MCP — beyond plain RAG 2026 edge · Module 11 →

Goal: build agents that use tools — the rarest skill on the market.

• The agent loop, tool calling, and structured outputs (Anthropic Messages API)
• Model Context Protocol (MCP): standard tools/resources across hosts
• Evaluating AI apps: golden sets, LLM-as-judge, prompt-injection defense

Interview hook: “agent vs RAG vs workflow, how you eval, how you stop prompt injection.”

Production Engineering — secure, fast, observable · Module 12 →

Goal: run it like it's real — not just make it work once.

• Security: OWASP, auth hardening, IDOR, injection, CORS, rate limiting
• Caching & background jobs: Redis, cache-aside, ARQ queues, idempotency
• Observability: structured logs, Sentry, OpenTelemetry traces, LLM cost

Interview hook: “how you'd debug a slow endpoint and secure a public API.”

Cloud & System Design · Module 13 →

Goal: deploy on AWS and pass the design interview.

• AWS for full-stack devs: IAM, S3, RDS+pgvector, Fargate/App Runner, secrets
• UAE data residency: the me-central-1 region
• System design framework: estimates, scaling, trade-offs — DocChat at 100k users

Interview hook: the whole system-design round, rehearsed on a system you actually built.