Solution Architect - Part-time - Worldwide

The Role

Part-time advisory role supporting our CTO on a multi-agent AI platform build. Sanity check architecture decisions, review designs, and guide critical technical choices across our 12-tool platform.

Stack: Next.js 14, TypeScript, Supabase, Railway, Anthropic Claude API, FFmpeg, Cloudflare R2

What You'll Advise On

• Agent orchestration: Multi-agent coordination patterns, task delegation, failure handling
• RAG architecture: Supabase pgvector implementation, embedding strategies, retrieval optimization
• Tool calling design: Agent tool interfaces, error handling, observability
• Data isolation: RLS patterns, multi-tenant security, data boundaries
• Queue & worker patterns: Background jobs, async processing, retry logic
• Phase 2 super agent: Morgan—unified conversational interface across all 12 tools
• Service boundaries: Where to split services, when to keep monolithic
• Observability: Logging, monitoring, agent audit trails
• Scalability: Performance bottlenecks, caching strategies, database optimization

Required Experience

Must Have:

• 10+ years shipping production systems at scale
• Shipped at least one multi-agent or LLM-orchestrated platform in production (not research, not prototypes)
• Deep distributed systems knowledge (microservices, event-driven architecture, eventual consistency)
• Production experience with LLM APIs (Anthropic Claude, OpenAI, or similar)
• Vector database experience (pgvector, Pinecone, Weaviate, Qdrant)
• Strong opinions on agent memory, audit logging, and observability
• PostgreSQL/Supabase expertise (RLS, query optimization, schema design)

Strongly Preferred:

• Multi-tenant SaaS architecture
• Next.js/TypeScript at scale
• Railway or similar PaaS
• RAG system design and optimization
• Video processing pipelines (FFmpeg)
• Queue systems (BullMQ, Inngest, SQS)