About

We are in the final hours of a game hackathon. We need an expert Prompt Engineer available right now to help craft/refine the System Prompt for our local LLM (Llama 3.1 8B Instruct GGUF running via LLMUnity/LLamaSharp in Unity). The LLM acts as an in-game "Judge" and must: - Understand player requests (specific game abilities) based on context provided each turn . - Reference game rules (costs, universal laws) provided in the prompt . - Engage in brief conversational clarification if needed. - Crucially: Reliably output structured JSON matching a defined schema. - Bonus: Experience with GBNF grammar for enforcing JSON output with llama.cpp/LLamaSharp is a major plus. - Deliverable: Optimized System Prompt string (and ideally a GBNF grammar file) + quick collaboration/testing with our developer via chat/call. Requirements: Proven prompt engineering expertise, especially with instruction-following and JSON output. Experience with Llama 3 models preferred. Must be available to start immediately and work efficiently for the next few hours. Please apply only if you are available right now and confident you can deliver quickly. Include relevant examples of forcing structured output from LLMs.