Yuna is building a conversational AI system for mental health support with a heavy ML/LLM stack (Python, PyTorch, JAX, TensorFlow, LangGraph) paired with enterprise infrastructure (Kubernetes, gRPC, WebRTC via LiveKit). The project list reveals a company mid-pivot toward B2B2C: they're shipping core conversational intelligence and self-harm detection safety layers while simultaneously investing in HRIS integrations, SSO, and eligibility file pipelines—signals of enterprise customer scaling beyond direct-to-consumer delivery.
Yuna operates a conversational AI platform designed to reduce barriers to mental health care access. The platform delivers 24/7 support through text-based interactions and integrates with enterprise systems (HRIS, identity providers, SSO). Core technical focus spans conversational quality evaluation, safety systems (self-harm detection), and scalable data pipelines. The company is headquartered in San Francisco and employs 11–50 people, with engineering and sales teams distributed across the United States, Ukraine, Mexico, Argentina, and Brazil. Active challenges include enterprise onboarding speed and reliability, architecture modernization, and data flow scalability.
Yuna's stack spans Python, Go, Node.js, LangGraph, Kubernetes, PyTorch, JAX, and TensorFlow for ML/conversational work, with gRPC and WebRTC (LiveKit) for real-time communication. Retool, Zapier, and n8n handle integrations and automation.
A core platform for conversational mental health support with embedded safety systems (self-harm detection), plus enterprise connectors: HRIS integrations, SSO, directory sync, and eligibility file pipelines. Current focus is improving onboarding reliability and scaling data flows.
Other companies in the same industry, closest in size