Senior Python Developer (AI, React, Java)

EPAM Systems; United States; Remote; USD 140,000 - 200,000; Full time

Job summary

EPAM Systems is seeking a Senior Python Developer to design, build, and deploy scalable intelligent agent infrastructure. You will focus on robust backend agent workflows, streaming protocols, and GitOps-driven deployment at scale.

The role requires deep asynchronous Python experience, LangGraph/LangChain expertise, and production-ready integration with CopilotKit and MCP tool interfaces. Expect Docker/Kubernetes environments and ArgoCD-based deployments.

Qualifications

• Expert in Python 3.11+ with async programming (asyncio, uvicorn).
• Hands-on experience with multi-agent systems using LangGraph and LangChain.
• Proven FastAPI experience (WebSockets, Background Tasks, SSE).
• Experience with CopilotKit Python SDK/runtime for co-agent interactions.
• Knowledge of MCP (Model Context Protocol) for tool integration servers.
• Docker and Kubernetes deployment; GitOps with ArgoCD/Helm.
• Nice to have 1+ year Java.
• Experience with vector databases (Pinecone, Milvus, Qdrant, PGVector).
• Frontend integration concepts (React/Next.js) and CopilotKit hooks.

Responsibilities

• Design, optimize, and maintain multi-agent state machines and tool-calling systems using FastAPI and LangGraph.
• Orchestrate model access, prompt routing, and vector store integrations with LangChain.
• Implement and scale the CopilotKit Python runtime for real-time backend/frontend streaming.
• Build standard interfaces for tool/data integrations using MCP for external service connections.
• Package services into Docker containers and deploy on Kubernetes with Helm and ArgoCD.
• Write high-performance, asynchronous Python code to handle parallel tasks and high-concurrency streaming.

Skills

Python

Async programming

FastAPI

LangGraph

LangChain

CopilotKit

MCP

Tools

Docker

Kubernetes

ArgoCD

Helm

LangSmith

Phoenix

Arize

Pinecone

Milvus

Job description

We are looking for a highly skilled Senior Python Developer . In this role, you will design, build, and deploy the next generation of our intelligent agent infrastructure. You will focus on building robust backend agent workflows, setting up high-performance streaming protocols for user-agent interaction, and ensuring seamless deployment at scale using GitOps. Our ideal candidate is passionate about the evolving AI ecosystem, deeply comfortable with asynchronous programming in Python, and experienced in bringing complex agentic workflows into production.

Responsibilities

• Agent Backend Development: Design, optimize, and maintain complex multi-agent state machines, graphs, and tool-calling systems using FastAPI and LangGraph
• AI Model Integration: Orchestrate and configure model access, prompt routing, and vector store integrations using the LangChain ecosystem
• Agent-to-UI Streaming: Implement and scale the CopilotKit Python runtime to establish bidirectional, real-time communication protocols between backend agents and frontend client components
• Integration & Interoperability: Build and implement standard interfaces for tool/data integrations using the Model Context Protocol (MCP) to allow external services to connect smoothly with our agent ecosystems
• Production Deployment: Package backend services into Docker containers and orchestrate deployments across Kubernetes clusters using Helm charts and GitOps CD engines like ArgoCD
• Performance & Reliability: Write clean, performant, and heavily asynchronous Python code (asyncio) capable of handling complex parallel tasks, high-concurrency streaming, and model callbacks

Requirements

• Core Language: Strong expertise in Python 3.11+, with deep knowledge of asynchronous programming (asyncio, uvicorn, anyio)
• AI/Agent Frameworks: Hands-on experience building stateful, multi-agent systems using LangGraph and utilizing LangChain for abstracting LLM runtimes, memory, and output parsing
• Proven experience with FastAPI (WebSockets, Background Tasks, SSE streaming)
• Direct experience or strong familiarity with the CopilotKit Python SDK / runtime for co-agent interactions
• Practical, working knowledge of the Model Context Protocol (MCP) for building standardized tool integration servers
• Docker containerization best practices (multi-stage builds, security pinning)
• Container orchestration with Kubernetes (writing/modifying Helm charts)
• Strong exposure to GitOps workflows and continuous deployment using ArgoCD
• Nice to have 1+ year of experience with Java
• Experience with vector databases (e.g., Pinecone, Milvus, Qdrant, PGVector)
• Understanding of frontend integration concepts (e.g., React, Next.js, and how CopilotKit hooks interact with the frontend components)
• Contributions to open-source AI frameworks (LangChain, LangGraph, CopilotKit, or MCP servers)
• Familiarity with distributed tracing for LLM applications (e.g., LangSmith, Phoenix, or Arize)