Senior Agentic AI Engineer: Build & Scale Autonomous Agents

EPAM Systems; Colombia; Remoto; COP 120.000.000 - 180.000.000; Tempo pieno

Vantaggi offerti da questo lavoro

Healthcare benefits

Paid time off

Descrizione del lavoro

EPAM Systems is seeking a Senior Agentic AI Engineer to design, build, and operate autonomous agent systems at scale for production deployments. You will drive architecture decisions across runtime, memory, identity, tools, and observability, leading the use of AWS Bedrock AgentCore and Strands SDK across agent harnesses.

You will design memory management, MCP servers, and tool integrations, aiming to optimize reliability, cost, and latency while enforcing guardrails and robust evaluation

Competenze

• 3+ years building production software.
• 1+ year shipping LLM-based agents to production.
• Proficient in Python and agent frameworks.
• Knowledge of production observability, guardrails, and security boundaries.
• English proficiency at B2+ level.

Mansioni

• Design the architecture for autonomous agents across runtime, memory, identity, tools and observability.
• Lead use of AWS Bedrock AgentCore and the Strands Agents SDK in deployments.
• Develop long-term memory and context-management systems for multi-step engagements.
• Build MCP servers and tool integrations for agent calls (recon, scanners, primitives, internal services).
• Improve reliability, cost and latency of agent reasoning and establish evals, tracing and guardrails.

Conoscenze

Python

Agent frameworks

LLM agents

Observability

Security boundaries

Strumenti

Strands

LangGraph

Claude Agent SDK

OpenAI Agents SDK

CrewAI

Descrizione del lavoro

EPAM Systems is seeking a Senior Agentic AI Engineer to design, build, and operate autonomous agent systems at scale for production deployments. You will drive architecture decisions across runtime, memory, identity, tools, and observability, leading the use of AWS Bedrock AgentCore and Strands SDK across agent harnesses.

You will design memory management, MCP servers, and tool integrations, aiming to optimize reliability, cost, and latency while enforcing guardrails and robust evaluation