SWE Interview Prep
Browse the public curriculum
— 18 tracks, 222 concepts, and 24 sequenced roadmaps without requiring
JavaScript or sign-in.
Product: learn.significanthobbies.com
A comprehensive interview preparation platform for mastering DSA, Low-Level Design, System Design, and Behavioral interviews.
Current qualification and tasks
The drill journey now has executable grading/persistence evidence,
including reload and two repaired editor-resume regressions. Source and CI
checks do not qualify hosted account persistence.
Drill, artifact, and project edits now show local-only, pending, failed, or
synced status. Pending account edits survive reload and retry on reconnect or
with Retry sync. Account changes hide the prior account’s records and drafts.
Older unscoped progress remains available when signed out; it is not silently
attributed to whichever account signs in next.
Rollout prerequisite: apply additive D1 migration
0003_record_sync_receipts.sql before deploying the new handlers/client. No
remote migration or deployment has been performed. Concurrent-tab and
cross-device record recovery is replay-tested against the real handlers;
conflict resolution stays server-ordered last-writer-wins, and notes, mastery,
and ELO remain outside this repair. Hosted qualification remains #97.
The task reconciliation found no prior open Issues or PRs; no historical
product intention was closed as complete.
Scale and platform follow-ups
Scale and the optional platform learning guides are shipped in the Play and
Systems Labs surfaces. The Scale qualification record
documents the evidence and limits from issue #99.
Campaigns already include bounded customer churn; the optional work below is
stronger realism and qualification, not a missing release requirement:
- Add workload-specific correctness, recovery, and delivery objectives so
different valid architectures create distinct decisions.
- Qualify queue, stream, cache, CDN, search, and replica strategies in
dedicated scenarios and continue balance work without arbitrary mandatory
tool checklists.
- Preserve incident-specific lesson selections and improve incident history
beyond the current bounded journal.
- Reduce the recorded first-party Scale complexity debt (15 functions over the
general threshold, with maximum CCN 58 and 174 lines).
The optional Platform Lab guides still require the existing external
platform-lab checkout and private vault-common sibling chart. SWE does not
bundle that infrastructure or claim cluster evidence; a portable kit remains
future work until that dependency is replaced and verified.
Problem
Technical interview preparation is fragmented across multiple tools: LeetCode for coding, Excalidraw for diagrams, ChatGPT for hints, and Anki for spaced repetition. Switching between tools breaks flow and makes it hard to track progress holistically.
SWE Interview Prep consolidates everything into a single platform with integrated code execution, diagram drawing, AI assistance, and spaced repetition review.
Deployment & External Services
| Concern | Service |
|---|
| Hosting | Cloudflare Pages (swe-interview-prep) + Pages Functions backend (functions/api/[[path]].js) |
| Database | Cloudflare D1 |
| Auth | Google One Tap (JWT cookie) |
| File storage | Cloudflare R2 (swe-interview-prep-assets) — hosts the Go WASM interpreter |
| Analytics | PostHog (via local posthog-js wrapper) |
| AI | Vercel AI SDK via @ai-sdk/openai-compatible against a BYO OpenAI-compatible endpoint (per-request aiConfig or AI_* env). Dev bridge streams local claude/codex/gemini CLIs. |
| CI/CD | GitHub Actions — ci.yml on push/PR; deploy.yml is manual (workflow_dispatch) |
Features
- Interactive Code Editor - Write and run TypeScript (transpiled and executed in-browser via sucrase) or Go (client-side WASM interpreter, with an
/api/go-run → go.dev fallback) in Monaco Editor (VS Code engine)
- Visual Design Tool - Draw system architecture diagrams with Excalidraw integration
- BYO-Endpoint AI Hints - Get Socratic guidance without spoilers through any OpenAI-compatible endpoint (dev: local claude/codex/gemini CLIs)
- Spaced Repetition System - Concept mastery via FSRS (
ts-fsrs); Anki decks can be imported
- LeetCode Import - Fetch problems directly via LeetCode API
- Progress Tracking - Monitor mastery across 18 learning tracks spanning interview fundamentals, systems, AI-native engineering, developer tools, applications, and multimodal computing
- Pattern-Based Learning - Group problems by algorithmic patterns (sliding window, two pointers, etc.)
This is the beginning of the README. Read the rest at the original. Read the full README ↗
