Appearance
Introduction
Ediky is a placement-preparation platform built and maintained by Shyama and Ayushi, targeting FAANG-tier and top product-company (D.E. Shaw, Flipkart) interview preparation for engineering students. It covers DSA practice, competitive programming, company-style OA simulation, CS fundamentals, aptitude, and SQL — with a real code-execution backend, AI-assisted grading, and spaced-repetition revision, all in one product at ediky.com.
Ediky Workflow — this site — is the engineering documentation for that product. It is written so that a new developer can understand the entire ecosystem from these pages alone: every repository, every pipeline, every scheduled job, every API, and the reasoning behind each decision.
What Ediky is, in one paragraph
Ediky is a React 19 + Vite SPA served from Cloudflare Pages, backed by Firebase (Auth + Firestore) and a self-hosted Judge0 CE instance on Oracle Cloud's free tier for real multi-language code execution. There is deliberately no traditional backend server: Cloudflare Pages Functions handle request/response work (AI grading, code-run proxying, notes fetching), Cloudflare Workers handle scheduled work (background OA judging, event reminders), and Firestore security rules handle authorization. The most interesting engineering lives in how the content layer scales — the question bank (~1,400 questions across DSA, CP, CS, aptitude, DSA-technical, and SQL sections today, designed for 10k–20k) is served as sharded JSON indexes plus on-demand markdown bodies rather than being bundled into the app.
The four repositories
| Repository | Role | Documented in |
|---|---|---|
ediky-website | The product — SPA, Pages Functions, Workers, content, judging | The Website, Backend & Infra |
ediky_tools | The content factory — question generation, validation, and shipping pipelines | Ediky Tools |
Docy | Lecture-notes markdown, fetched at runtime through a Pages Function proxy | Docy |
ediky-core | This documentation portal (deployed as Ediky Workflow) | Development Setup |
All four repositories are private. That is why this portal carries no GitHub links or icons — they would be dead ends for anyone without repo access.
How to read these docs
- New to the project? Read this page, then Ecosystem & Repositories, then System Architecture. Those three give you the mental model everything else hangs off.
- Working on a feature? Jump straight to the relevant section — each Website page documents one subsystem end to end (UI → lib → function → data).
- Authoring questions? Everything you need is under Ediky Tools; the format rules that the linter enforces are summarized in Shared Utilities.
- Preparing for an interview about this project? Challenges & Tradeoffs and Interview Q&A are written exactly for that.
- Stuck? Use the Ask AI button (bottom-right). It answers from these pages and cites the sections it used — see AI Pipeline for how it works.
Scope, honestly
This is a project built by a small two-person team, not a funded startup with a large team. Some things are intentionally simple where a company product would invest more — there is no dedicated backend service, no Kubernetes, no message queue. Where a simplification carries a real tradeoff, it is discussed openly (see Challenges & Tradeoffs) rather than hidden. The documentation follows the same rule: it records what is, including the sharp edges.