Spatium Lab

RAG, LLM Integration, Function Calling, AWS Bedrock, MCP/A2A

Build AI-powered features for B2B SaaS that handle real customer data securely. This is practical AI engineering, not research—you'll ship production systems that integrate LLMs, implement RAG pipelines, and build agent workflows. We're looking for engineers who understand that AI in production means thinking about security, multi-tenancy, costs, and reliability, not just prompts.

• Design and implement RAG pipelines for engineering documentation and knowledge bases
• Build function/tool calling systems with structured outputs and reliable execution
• Integrate LLMs via AWS Bedrock, GCP Vertex AI, or OpenAI APIs
• Develop MCP (Model Context Protocol) servers and A2A (Agent-to-Agent) workflows
• Implement secure, multi-tenant AI features that protect customer data isolation
• Build background task pipelines (Celery, SQS) for async AI processing at scale
• Create embeddings pipelines and manage vector databases for semantic search
• Monitor AI system performance, costs, and quality metrics in production

Requirements

• Strong Python skills with async programming experience
• Understanding of transformer architectures and LLM fundamentals—how they work, not just how to call them
• Experience building with LangChain, LlamaIndex, or similar orchestration frameworks
• Knowledge of vector databases (Pinecone, Weaviate, Chroma, pgvector)
• Familiarity with AWS or GCP cloud services and containerization (Docker)
• Understanding of security principles for multi-tenant systems—data isolation matters
• Experience with message queues and background processing (Celery, SQS, Redis)
• Self-directed problem solver who designs solutions, not just implements specs
• Proficient with AI coding tools and prompt engineering—you've built with these, not just experimented
• Prior professional experience not required if portfolio demonstrates real AI engineering capability
• Competitive hourly compensation
• Flexible schedule (full-time summer or part-time year-round)
• Direct mentorship from founders and senior engineers
• Real project ownership with production impact
• Path to full-time conversion for exceptional performers
• Modern tech stack and cutting-edge AI tools
• Small team, high autonomy environment