Senior AI Systems Engineer — LLM Orchestration & Multi-Agent (Spiritual App) - Contract to Hire
JOB DESCRIPTION
We are building a Spanish-first AI-powered spiritual companion app.
This is NOT a simple chatbot.
We are building a multi-layer AI system with strict guardrails, theological consistency, and dynamic response protocols.
We are looking for a senior AI systems engineer who understands LLM architecture, not just prompt writing.
⸻
Scope of Work
You will design and implement the AI system that powers the app, including:
- Multi-layer prompt architecture (core + companion + situational protocols)
- Intelligent routing system (detect user intent and load correct protocols)
- Guardrails system (theological, ethical, safety)
- Response validation layer (post-generation filtering / correction)
- Multi-companion structure (different personalities + functions)
- Context handling (short-term memory + conversation flow)
- Voice + text integration (speech-to-text + text-to-speech ready)
⸻
⚠️ Important
This is a logic-driven AI system, not a UI task.
You are NOT just calling OpenAI.
You are designing how the system thinks, routes, and responds.
⸻
✅ Requirements
- Strong experience with OpenAI API (or similar LLMs)
- Experience building AI-powered applications (not prototypes only)
- Understanding of prompt architecture and system design
- Backend thinking (Python preferred)
- Experience with:
- LLM orchestration
- AI agents or multi-step flows
- Context management
- Ability to translate complex logic into stable systems
⸻
Not a fit if:
- You only have frontend / Flutter experience
- You only “use ChatGPT”
- You have not built real AI systems in production
⸻
Ideal Candidate
- Thinks in systems, not features
- Has built LLM-based apps end-to-end
- Understands trade-offs (cost, latency, quality)
- Can challenge ideas and improve architecture
⸻
To Apply
Please answer:
1. Describe an AI system you built (not just prompts — full architecture)
2. How would you design a system that uses different prompts depending on user intent?
3. How would you prevent incorrect or unsafe AI responses?
4. Have you worked with multi-agent or multi-layer AI systems?
⸻
Goal
We are building a high-quality, scalable product — not an experiment.
Apply Now
Apply Now