Principal Conversation & Agentic Experience Architect

Twin Health
•$220,000 - $240,000•Remote

About The Position

Twin Health is hiring a Staff-level architect to shape how its AI agents understand members, communicate, collaborate, personalize experiences, and escalate to human care. The role will define a coherent, trustworthy experience across the Digital Twin, specialist AI agents, clinical care teams, and human coaches, and translate behavioral principles into executable systems used by product, clinical, and AI teams. The architect will establish Twin’s principles for intelligent interactions and maintain its voice, tone, personality, and behavioral standards. They will design multi-agent roles, boundaries, handoffs, and context-preservation protocols; establish personalization, trust, safety, privacy, consent, transparency, and clinical-boundary frameworks; and partner with AI engineering to encode standards in reusable Skills, agent templates, prompts, and governed context/MCP integrations. The role also owns evaluation and regression testing, agent-behavior governance, and cross-functional alignment and enablement. This is a remote opportunity based in the U.S. Applicants must be authorized to work for any U.S. employer; visa sponsorship or transfer is not available. The listed annual compensation range is $220,000–$240,000.

Requirements

  • At least 8 years of experience in conversation design, content design or architecture, UX for AI, model behavior, or an adjacent field.
  • At least 3 years shaping LLM-based product experiences in production.
  • Staff-level scope defining behavior, voice, or interaction systems adopted across multiple teams or product surfaces.
  • Exceptional writing and judgment, including the ability to craft effective model responses and specifications that produce consistent responses at scale.
  • Hands-on fluency with modern LLM tooling, including prompt and system-prompt design, agent frameworks, tool use or MCP-style integrations, and evaluation design such as automated and human evaluations and rubrics.
  • Technical intuition for how prompts, context, and data affect model behavior; an ML engineering background is not required.
  • Systems-thinking experience designing multi-agent, multi-surface, or multi-channel experiences where context, state, and persona persist across boundaries.
  • Ability to translate ambiguous questions about autonomy, trust, and healthy human-AI interaction into concrete, testable standards.
  • Proven cross-functional leadership aligning research, engineering, product, design, and domain experts, including clinicians.
  • Authorization to work for any employer in the U.S.; Twin cannot sponsor or take over sponsorship of an employment visa.

Nice To Haves

  • Experience in healthcare, digital health, coaching, or another regulated or high-trust domain; HIPAA familiarity is a plus.
  • Experience with safety-critical model behavior, including escalation design, crisis pathways, clinical boundaries, and refusal or deferral patterns.
  • Familiarity with Anthropic’s Claude ecosystem (Skills, MCP, agent SDKs) or equivalent tools such as OpenAI Assistants/Agents or Gemini.
  • Background in linguistics, HCI, behavioral science, service design, or content strategy.
  • Experience building or operating LLM evaluation infrastructure and data flywheels using production signals.

Responsibilities

  • Define the principles governing intelligent interactions across the Twin ecosystem, including health, motivation, personalization, autonomy, encouragement, intervention, and human care.
  • Translate the interaction philosophy into behavioral specifications that can be built, evaluated, and enforced.
  • Design and maintain a living, versioned voice, tone, and personality system covering empathy, encouragement, explanation, insights, goal setting, intervention, and escalation.
  • Ensure consistency between agents so members experience a coherent service.
  • Map specialist-agent roles, responsibilities, boundaries, context requirements, and overlaps.
  • Define when agents act, listen, hand off to specialists, or bring a coach or clinician into the loop; design context-preservation protocols for seamless transitions.
  • Define how interactions adapt to member goals, state, history, progress, and preferences.
  • Establish standards for safety, privacy, consent, clinical boundaries, transparency, and healthy AI interaction, including safeguards against dependence, confusion, and over-engagement.
  • Partner with AI engineering to encode standards as reusable Claude Skills, specialist-agent templates, system prompts, and governed context/MCP integrations.
  • Build and own an evaluation and regression suite that tests whether agents remain aligned with Twin’s principles as models, prompts, and experiences evolve.
  • Establish governance for creating, reviewing, versioning, and approving agent behaviors and Skills.
  • Align product, design, content, clinical, engineering, and leadership teams around shared standards without direct authority.
  • Run workshops and enablement to help teams create specialist experiences using Twin’s foundational intelligence.
  • Perform other duties as assigned.

Benefits

  • Competitive compensation package in line with leading technology companies.
  • Opportunity for equity participation.
  • Unlimited vacation with manager approval.
  • 16 weeks of 100% paid parental leave for delivering parents and 8 weeks for non-delivering parents.
  • 100% employer-sponsored healthcare, dental, and vision coverage for the employee, and 80% coverage for family.
  • Health Savings Account and Flexible Spending Account options.
  • 401(k) retirement savings plan.
  • Remote work and flexibility.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service