Built an AI-native mental health companion focused on proactive support, behavior change, and provider collaboration. Designed around memory, agentic workflows, trust, escalation, and healthcare-adjacent safety.
Context: Cairah was built to help people manage stress, anxiety, and depression more proactively, while creating better continuity, earlier support, and smarter navigation toward care.
Revenue / Portfolio Context: Cairah was an early stage founder led AI product built to MVP and launch stage as a customer facing AI-native platform. Alongside the product, we also built partnerships with therapists across the East Coast, creating an early ecosystem around care navigation, referral support, and behavioral health collaboration.
Problem: Most mental health tools are too reactive, too generic, or too disconnected from the user's real context over time. We wanted to build something more continuous, supportive, and personalized without overstepping trust or safety boundaries.
What I Did: I led the product from concept to MVP, including:
- Conversation design
- Behavior-change workflows
- Memory strategy
- Escalation and severity handling
- Provider collaboration design
- AI architecture decisions
AI / System Design: The system used:
- LangGraph for agentic orchestration
- LangSmith for tracing and evaluation visibility
- Postgres + pgvector for memory retrieval
- Embeddings to connect past context to current conversation
UX / Workflow Thinking:
- Understand the user over time
- Guide better next steps
- Escalate when needed
- Create trust through continuity and boundaries
Business Impact:
- Launched a live MVP in the App Store
- Validated a meaningful behavioral health use case
- Created a strong AI-native product foundation around trust, memory, and escalation
- Generated early product and market learning for future scale