Lead Autonomous Agent Engineer for Production-Grade AI

EPAM Systems; Colombia; Remote; COP 180,000,000 - 300,000,000; Full time

Job summary

EPAM Systems seeks a Chief Agentic AI Engineer to design and scale autonomous agents across runtime, memory, identity, tools, and observability. You will lead production-grade deployments using AWS Bedrock AgentCore and agentic frameworks, shaping guardrails, evals, and reliability.

You will architect long-term memory and context-management, develop MCP servers and tool integrations, and optimize performance and cost of agent reasoning across multi-step engagements.

Qualifications

• 7+ years building production software with 2+ years shipping LLM-based agents.
• Expert Python and hands-on use of agent frameworks (Strands, LangGraph, OpenAI Agents SDK, etc.).
• Experience with AWS Bedrock AgentCore primitives and production observability.
• Deep understanding of memory architectures, tool use, and agent evaluation.
• Design for production: guardrails, retries, cost control, and security boundaries.
• English at B2 level or higher.

Responsibilities

• Define architecture for autonomous agents across runtime, memory, identity, tools, and observability.
• Lead deployments using AWS Bedrock AgentCore and agent harnesses.
• Architect long-term memory and context-management for multi-step engagements.
• Develop MCP servers and tool integrations, including recon and scanners.
• Optimize reliability, cost, and latency of agent reasoning.
• Establish evals, tracing, and guardrails standards.

Skills

Python

Agent Frameworks

AWS Bedrock

Memory architectures

Observability

English proficiency

LLM-based agent deployment

Tools

LangGraph

Claude Agent SDK

OpenAI Agents SDK

CrewAI

Strands

Job description

EPAM Systems seeks a Chief Agentic AI Engineer to design and scale autonomous agents across runtime, memory, identity, tools, and observability. You will lead production-grade deployments using AWS Bedrock AgentCore and agentic frameworks, shaping guardrails, evals, and reliability.

You will architect long-term memory and context-management, develop MCP servers and tool integrations, and optimize performance and cost of agent reasoning across multi-step engagements.