Description: * Solid Python (asynchronous, typing, Pydantic or similar). * Solid PHP knowledge for integration with the legacy platform. * LangChain (runnables/tools/callbacks). * PostgreSQL (queries, modeling, and performance optimization). * MCP (Model Context Protocol) — understanding of clients/servers and integration with agents. Advantages * End-to-end RAG (ingestion, chunking, retrieval, and evaluation). * n8n (workflows, credentials, error handling, automations). * Vector databases (e.g., Chroma, pgvector, Milvus, Pinecone) and best practices for indexing/recall. * Agent workflow engines/orchestration (e.g., state machines, checkpoints). * LLM observability/tracing (e.g., LangSmith, OpenTelemetry) and dashboards (e.g., Grafana). * SRE practices in production (metrics, logs, alerts) and structured logging. * Integration with external services (e.g., BaaS, third-party APIs) and asynchronous HTTP clients. * Design and implement multi-LLM agent and tool pipelines (e.g., OpenAI, Anthropic, Google). * Build and maintain API services and asynchronous integrations (modern web frameworks e.g., FastAPI/Flask/Django). * Data modeling and persistence in PostgreSQL; schema migrations and versioning. * LLM observability (tracing/metrics/logs) and performance instrumentation. * Automated testing and continuous quality assurance (linters/formatters/pytest). * API security (authentication, tokens/JWT, rate limiting, CORS) and containerized deployment (e.g., Docker/K8s). Your mission will be to evolve and maintain production LLM agent APIs and workflows, focusing on end-to-end quality, observability, and efficiency. 2512060202191761547