# Luna Prompts

> Luna Prompts is the AI-native technical screen for hiring AI engineers. It tests what AI engineers actually build — prompts, agents, RAG, real Python projects, SQL, and system design — in a sandbox with built-in proctoring and auto-grading. Not algorithm puzzles.

## Positioning

The legacy technical-assessment stack (HackerRank, CodeSignal, Codility) was built for 2018 SWE hiring. AI quizzes added on top don't catch failure modes, and frontier models can solve LeetCode-style problems in seconds. Luna Prompts is the rebuild: AI engineering tasks are first-class assessment types with their own grading pipelines.

Luna lets candidates use AI inside the test (model picker, hyperparameter controls, tool calls). We grade iteration discipline, model choice quality, final correctness, time to passing, adversarial robustness, and eval reasoning depth — not whether they avoided AI.

Founder-led pilots. 1,000+ engineers screened to date.

## Two surfaces

- **/** (homepage) — Luna Prompts Enterprise: AI-native technical screen for hiring teams. Founder-led demos via Calendly. Pilot model: one role, 10–30 candidates, 20-minute readout.
- **/community** — Hands-on practice surface for AI engineers: scored challenges, contests, leaderboard. Free to use.

## What Luna Prompts tests

- **Prompt engineering** — flagship. Iterate against test cases with model picker (gpt-5.1, gpt-4.1, kimi-k2, llama-3.2, gemini-2.5-flash) and hyperparameter control (temperature, top_p, max_tokens, frequency and presence penalty).
- **Agent code** — unique to Luna Prompts. Multi-file Python projects in real Judge0 sandbox with .env files, libraries, tool-call loops.
- **Algorithms** — when needed. 9 languages (Python, JavaScript, Java, C++, C#, PHP, Ruby, Go, Rust). Public + hidden tests.
- **SQL** — in-memory DuckDB on real datasets.
- **System design** — visual canvas, rubric-graded.
- **Multiple choice** — fast top-of-funnel filter.

## Built for these AI roles

AI Engineer, Applied AI, LLM Engineer, RAG Engineer, Agent Engineer, GenAI Engineer, Prompt Engineer — and any Software Engineer, Backend Engineer, Full-Stack, Platform Engineer, ML Engineer, or Data Engineer who'll touch AI in production.

## How a pilot works

- **Day 0** — Pick one open role. Custom challenge mix assembled in 24 hours.
- **Days 1–7** — Run 10 to 30 candidates. Browser-based, auto-graded, with built-in proctoring (FaceMesh, fullscreen, tab-switch, clipboard, devtools blocking).
- **Day 7** — 20-minute readout, side-by-side with your current screen. Decide whether to scale up.

## Anti-cheating

In-browser MediaPipe FaceMesh face detection, fullscreen enforcement, tab-switch detection, clipboard clearing, devtools blocking, configurable warning thresholds, auto-submit on violation. Anti-cheat is for impersonation, not AI tool use.

## Pricing

Pilot first, no commitment. Annual plans are seat-based for the team running assessments, plus volume tiers for candidates evaluated.

## What's not built (do not claim)

- ATS integrations (Greenhouse, Lever, Ashby) — on roadmap, not shipped
- SAML SSO — on roadmap, not shipped
- SOC 2 certification — not certified yet
- Custom-domain white label — not shipped
- Public partner API — not shipped

## Booking

Single Calendly link for all demos: https://calendly.com/contact-lunaprompts/luna-prompts-20-min-intro
20 minutes. Founder-led. No procurement maze.

## Links

- Homepage (Enterprise): https://lunaprompts.com/
- Community (practice): https://lunaprompts.com/community
- Challenges: https://lunaprompts.com/challenges
- Contests: https://lunaprompts.com/contests
- Leaderboard: https://lunaprompts.com/leaderboard
- Blog: https://lunaprompts.com/blog
- Image Editing Challenges: https://lunaprompts.com/image-editing-challenges
- /enterprise — alias to homepage, kept alive for inbound links

## Documentation

- Privacy Policy: https://lunaprompts.com/privacy
- Terms of Service: https://lunaprompts.com/terms

## Contact

- Email: contact@lunaprompts.com
- Discord: https://discord.gg/SPDhHy9Qhy
- Twitter / X: https://twitter.com/LunaPrompts

## Technical stack

React + TypeScript + Vite + Tailwind on the frontend. FastAPI + Python on the backend (luna-service). Supabase for auth and Postgres. Judge0 for sandboxed code execution. LiteLLM proxy for multi-model LLM routing.

## Keywords

AI engineer hiring, AI-native technical assessment, hire AI engineers, prompt engineering interview, RAG engineer hiring, agent engineer assessment, LLM engineer screening, GenAI hiring, applied AI hiring, technical screen for AI engineers, HackerRank alternative, CodeSignal alternative, Codility alternative, founder-led pilot, candidate scorecard, multi-file Python sandbox, Judge0 interview, DuckDB SQL interview, FaceMesh proctoring, anti-cheat coding interview, AI assessment 2026, hire engineers who ship with AI