Python Agentic AI Developer

We are looking for a skilled AI/ML Developer to join our agentic platform team on a contractual basis. You will design, develop, and deploy LLM-powered agents and pipelines that directly serve field technicians in production. You will work closely with the platform architect and contribute to the core multi-agent orchestration system. Key Responsibilities

• Design and develop LLM-powered agents using LangGraph and LangChain frameworks
• Build and enhance multi-agent orchestration pipelines with conditional routing, state management, and error recovery
• Develop RAG pipelines with hybrid search (semantic + keyword), re-ranking, and Reciprocal Rank Fusion
• Implement tool-calling agents with parallel API execution, sequential enrichment chains, and summarization
• Build real-time streaming capabilities (WebSocket / SSE) for conversational AI
• Integrate with Azure OpenAI APIs with circuit breaker patterns and fallback chains
• Implement prompt engineering and dynamic prompt management (DB-backed with in-memory caching)
• Design and implement multi-modal AI features (image analysis with vision models)
• Write production-quality Python code with error handling, logging, observability, and unit tests
• Participate in code reviews, architecture discussions, and sprint delivery

Must-Have

• 10+ years of software development with strong Python expertise in production environments
• 2+ years hands-on LLM development (LangChain, LangGraph, or equivalent frameworks)
• RAG pipeline experience - embeddings, vector search, re-ranking, answer generation
• LLM APIs - OpenAI Chat Completions, Responses API, streaming, tool calling
• Async Python - asyncio, httpx, FastAPI
• Prompt engineering - system prompts, few-shot, structured JSON output
• Circuit breaker patterns, retry logic, resilient service design
• PostgreSQL and async ORMs (SQLAlchemy async / Alembic)
• Git, CI/CD, Docker, Kubernetes basics
• Strong system design skills - microservices, distributed systems, API design

Nice-to-Have

• LangGraph StateGraph (conditional edges, node pipelines, compiled graphs)
• Multi-agent systems - orchestration, routing, RBAC-based access control
• Langfuse or similar LLM observability (tracing, token usage, latency)
• WebSocket and SSE streaming protocols
• Multi-modal AI (vision models, image analysis)
• Vector databases (OpenSearch, Pinecone, pgvector)
• AWS services (Aurora PostgreSQL, SSM, IAM, EKS)
• Telecommunications or field service domain experience

Tech Stack

Category Technologies Language Python 3.11+ Frameworks FastAPI, LangChain, LangGraph LLM APIs Azure OpenAI (GPT-4) Databases PostgreSQL (Aurora), OpenSearch (vector) Messaging WebSocket, SSE, REST Observability Langfuse, Datadog, Splunk Infrastructure Docker, Kubernetes (EKS), AWS CI/CD GitLab CI, Helm, ArgoCD Package Mgmt uv (Python), pyproject.toml Testing pytest, unittest, FastAPI TestClient

The TechMobile Agentic AI team builds and operates a production-grade multi-agent AI platform serving 40,000+ daily interactions for field service technicians across a major telecommunications provider. The platform uses LLM-powered agents to provide intelligent troubleshooting, equipment lookups, job preparation, and visual inspection through a conversational natural language interface. Our tech stack is modern and we work on real AI systems at enterprise scale, not prototypes., AgreeYa is a global systems integrator delivering competitive advantage for its customers through software, solutions, and services. Established in 1999, AgreeYa is headquartered in Folsom, California, with a global footprint and a team of more than 1,850 staff across offices. AgreeYa works with 500+ organizations ranging from Fortune 100 firms to small and large businesses across industries such as Telecom, Banking, Financial Services & Insurance, Healthcare, Utility & Energy, Technology, Public sector, Pharma & Biotech, and others. Please visit us at Agreeya for more information.