Read this first
This programme isn't tutorials-and-quizzes. Every single day you build something that actually runs. Over 8 weeks these small pieces add up to a live job platform used by real people in Pakistan.
You'll be trained as an FDE, not a "backend dev" or "frontend dev".
🚀 FDE = Forward Deployed Engineer. It means you own a real problem end to end — you touch the whole system (scraping, backend, frontend, deploy) and do whatever engineering it takes to get a working thing into a user's hands. You own the outcome, not just your ticket.
Honest expectation: this is experience-focused. You walk away with a standout portfolio project and real skills — it is not a guaranteed permanent job.
The 8 weeks
Week 1
Get dangerous across the whole stack
One week to go from "knows Python" to "can build a full app." Every day ends with a working, committed piece.
You'll build
A small full app: a Next.js page talking to a FastAPI backend, saving to a Postgres database, filled with data you scraped yourself.
You'll learn
Proper Python (types, async), building APIs with FastAPI, databases, first taste of scraping, and a Next.js frontend.
You'll show
The full loop running end-to-end by Friday.
Week 2
Build a scraper framework (not 50 scripts)
Instead of writing throwaway code for every site, you design one clean system that any new site plugs into.
You'll build
A reusable scraper "template" and one standard shape for a job, so 10 different sites all come out looking the same. Adding site #51 becomes a quick config, not a rewrite.
You'll learn
How to design software so it scales — contracts, data models, keeping raw vs cleaned data.
You'll show
10 platforms feeding into one clean, consistent format.
Week 3
Scale to 50+ sites & beat the ones that fight back
Push past 50 sources, including tricky sites that try to block bots. You learn that scraping at scale is about reliability, not just parsing.
You'll build
A robust system with retries, rate-limits and health checks — plus de-duplication so the same job on 5 sites becomes one clean record.
You'll learn
Handling failures gracefully, browser automation for hard sites, spotting when a source breaks.
You'll show
50+ sources running reliably, with clean de-duplicated data.
Week 4
Put it on servers & open an API
Move the engine off your laptop and make it real infrastructure — scheduled, containerised, and reachable by the website.
You'll build
Scheduled background jobs (one per site) running on the Codeaza server, plus a clean read API with search and filters that Rozgar's frontend can call.
You'll learn
Job scheduling, Docker, deploying, and designing an API for a real frontend.
You'll show
The data engine running on its own, serving jobs through an API.
Week 5
Build the job feed people actually use
Build the heart of rozgar.codeaza.org — a fast, searchable job feed that works great on a phone.
You'll build
The job feed: fast search, sensible filters, infinite scroll, clean job pages — all mobile-first (most users are on slow phones).
You'll learn
Frontend that feels fast, real loading/empty/error states, and product judgement (what a job card should actually show).
You'll show
A working, good-looking feed running off your Week-4 API.
Week 6
Accounts & job alerts
The feature that makes Rozgar sticky: a user saves a search and gets pinged the moment a matching job appears.
You'll build
Login/accounts, saved searches, and an alert engine that emails/WhatsApps users when a new matching job is scraped — without spamming them.
You'll learn
Authentication, matching logic, and closing the loop from scraper → user notification.
You'll show
A real user getting a real alert for a real job.
Week 7
Make it fast, tested & monitored
Real products don't fall over. Harden Rozgar so it's quick, trustworthy, and you can see its health at a glance.
You'll build
Caching for sub-second search, error tracking (Sentry) across the app, tests on the important paths, CI on every change, and a dashboard showing pipeline health.
You'll learn
Performance tuning, testing, monitoring — the "Codeaza standard" for production work.
You'll show
A fast, monitored, tested product with a health dashboard.
Week 8
Launch, own a number, defend your work
Ship to real users, pick one metric you own and improve it, then present your 8 weeks like an engineer defending real work.
You'll build
Your features live in production, plus a metric you moved (with before/after numbers) and a clean repo + runbook.
You'll learn
Ownership, measuring impact, and communicating your work.
You'll show
A 20-minute demo + architecture walkthrough to the whole team.
How we work
The ground rules
- Ship something running every day — no passive watching of tutorials.
- You'll present your work at the end of each phase (a short demo), and defend the whole build in Week 8.
- Muhammad Asim is your Engineering Manager for this programme — technical questions and direction come to him.
- Mubeen handles your onboarding, access, and day-to-day coordination (standups, repos, setup).
- The goal is a successful internship: you leave with a live product and real experience — the kind of portfolio piece most engineers never get.
FDE — Forward Deployed Engineer (owns a problem end-to-end)
API — how the website asks the backend for data
Scraper — code that collects data from websites
De-dup — removing repeated copies of the same job
CI — automatic checks on every code change