AI Systems Engineer (Multi-Agent & Conversational Intelligence)

We are building a consumer-facing mobile product that blends habit formation, interactive systems, and AI-driven conversational experiences. The product uses a multi-agent AI architecture, where distinct AI "characters" (each with defined roles, tone, and expertise boundaries) guide users through different domains of the experience. We're looking for an AI Systems Engineer to own the design, reliability, and evolution of this conversational layer. This role is NOT for someone who wants to quickly build another chatbot. This role is for someone who is looking for a slightly crazy opportunity at the edge of innovation, working with an extensively-connected, world-renowned team of neurologists, marketers, public health advocates, designers, and developers. CORE RESPONSIBILITIES MULTI-AGENT ROUTING AND ORCHESTRATION Design and maintain logic that routes user queries to the appropriate AI agent based on: - Topic detection - Direct addressing - Conversation continuity Ensure agents stay within clearly defined expertise boundaries Implement smooth, explainable handoffs between agents when topics shift Prevent "agent drift" or overlapping responsibilities AGENT BEHAVIOUR, TONE AND GUARDRAILS Implement and refine: - Agent-specific personalities - Tone and response constraints - Few-shot prompting patterns Ensure responses are: - Supportive, accurate, and grounded - Non-diagnostic and non-alarmist - Consistent with each agent's intended role Balance expressiveness with reliability and safety KNOWLEDGE RETRIEVAL AND CONTEXTUAL ANSWERING Own the retrieval pipeline for expert reference materials (e.g. documents, guides, research summaries) Ensure retrieved information is: - Relevant - Used selectively (not dumped verbatim) - Appropriately contextualised Reduce hallucinations through retrieval discipline and prompt design Support real-time updates to the knowledge base without redeploying code CROSS-FUNCTIONAL COLLABORATION You will collaborate with: - Mobile engineers (to expose AI capabilities cleanly in the app) - Interactive / game developers (to trigger AI responses from system events) - Product and content teams (who manage agent configuration and reference materials) REQUIRED SKILLS Strong experience building LLM-powered systems beyond basic chat Hands-on experience with: - Multi-agent or tool-routed AI systems - Prompt engineering with guardrails - Retrieval-augmented generation (RAG) Deep understanding of: - Context window tradeoffs - Hallucination mitigation - AI safety in user-facing products Ability to explain technical decisions clearly to non-technical teammates Pragmatic, product-minded approach (clarity and trust over novelty) BONUS SKILLS Experience with: - Conversational UX - Voice/text-to-speech systems - Health-, education-, or behaviour-change products Familiarity with game-adjacent or interactive AI use cases Strong opinions on AI reliability and user trust (with the ability to justify them) WHAT SUCCESS LOOKS LIKE After the first 1-2 months: - Agent routing is predictable, explainable, and testable - Responses feel intentional and bounded, not "chatbot-like" - Content updates can be made safely without breaking behaviour - Failures are obvious and controlled, not silent After 3-4 months: - The AI layer becomes a durable, extensible product asset - New agents, features, and tiers can be added without re-architecture - AI costs and performance are predictable - Users trust the system and use it every day