Agentic AI Engineer (Google ADK / GCP)

2 months ago

Bogotá, ColombiaMid Level

Responsibilities

Architect and implement multi-agent systems using sequential pipelines, parallel fan-out/gather patterns, and loop-based self-correction with Google ADK.
Build deterministic graph workflows that combine adaptive AI reasoning with explicit execution paths.
Create and integrate custom agent skills, third-party tools, Google Maps MCP, search tools, and enterprise APIs.
Evaluate and debug agent execution trajectories, loop limits, key collisions, hallucinations, and tool-calling behavior.
Deploy optimized agents to Agent Engine through Cloud Run and Google Cloud Platform while maintaining availability, security, and low latency.
Collaborate with product managers and AI researchers to optimize Gemini model implementations within agentic frameworks.

Bachelor’s degree in Computer Science, a related technical field, or equivalent practical experience.
At least 2 years of experience building and deploying production-grade LLM applications or agentic AI systems.
Strong proficiency in Python or TypeScript/Node.js, including Python environment management with uv or pip.
Experience with Google Cloud Platform, including Cloud Run, Vertex AI, Secret Manager, and Cloud Storage.
Hands-on experience with the open-source Google Agent Development Kit, version 2.0 or later, is preferred.
Understanding of multi-agent orchestration patterns, state graph architectures, and deterministic routing is preferred.
Familiarity with Model Context Protocol and external tool integration is preferred.
Experience implementing CI/CD pipelines for AI applications, including GitHub Actions or Terraform, is preferred.
Background in AI-agent evaluation frameworks and benchmarking precision, recall, and tool-calling accuracy is preferred.

1,001-5,000 employees