AI Engineer - Agentic Systems

Toptal; United States; Remote; USD 180,000 - 260,000; Full time

Job summary

Toptal is seeking a hands-on AI Engineer to design, build, and deploy production-grade agentic systems for enterprise environments. The role focuses on multi-agent orchestration, tool use, and dynamic context management.

You will implement safety guardrails, integrate frontier models, and work with leading agent frameworks on cloud infrastructure. This is a remote US-based position with preference for East Coast time and minimal travel.

Qualifications

• Hands-on experience shipping autonomous agent systems in production.
• Proficiency with model SDKs: Anthropic Claude Agent, OpenAI Agents, or Google Vertex AI.
• Experience deploying distributed apps on Google Cloud Platform (GKE, Cloud Run, Pub/Sub, BigQuery).
• Strong API design, data validation (JSON/Pydantic) and microservices architecture.
• Based in the US (East Coast time preferred).

Responsibilities

• Design, build, and deploy production-grade AI agents capable of multi-agent orchestration and tool-calling.
• Implement safety guardrails, context management, and robust error handling.
• Integrate external APIs and services via standardized protocols.
• Scale and operate agentic microservices on cloud infrastructure.
• Document interfaces and ensure reliable production deployments.

Skills

Agent development

Production engineering

Model SDK

Cloud GCP

API design

Microservices

JSON schemas

Pydantic

Job description

Summary

We are seeking a highly hands-on AI Engineer with a proven track record of shipping agentic AI systems into production for enterprise environments. In this role, you will design, build, and deploy production-grade AI agents capable of multi-agent orchestration, complex tool execution, and dynamic context management. You will implement robust safety guardrails and integrate frontier models using leading agent frameworks and cloud infrastructure. This position is remote within the US (East Coast time zone slightly preferred) with minimal travel required.

Tasks and Deliverables

• Design and deploy autonomous and multi-agent orchestration workflows capable of executing complex, multi-step enterprise processes.
• Implement structured tool-use frameworks and integrate external APIs and services using protocols such as Model Context Protocol (MCP).
• Engineer advanced prompt strategies, dynamic context management, and memory layers to optimize model performance and reasoning accuracy.
• Implement production-grade guardrails, schema enforcement, latency optimization, and defensive exception handling across all agentic workflows.
• Deploy, scale, and maintain agentic microservices on cloud infrastructure using modern containerized and serverless platforms.

Required Experience

• Production Agent Experience: Demonstrated, hands-on software engineering experience building, testing, and shipping autonomous agent systems and tool-calling workflows into production environments.
• Framework & Model SDK Expertise: Strong proficiency in at least one major model ecosystem: Anthropic (Claude Agent SDK, Model Context Protocol), OpenAI (OpenAI Agents SDK, Responses API, File Search), or GCP Vertex AI (Gemini, Google ADK / Vertex AI Agent Engine).
• Cloud Infrastructure (GCP): Experience building and deploying distributed applications on Google Cloud Platform, using services such as GKE, Cloud Run, Pub/Sub, and BigQuery.
• Core Software Engineering: Strong foundation in software design patterns, API integrations, structured outputs (Pydantic/JSON schemas), and microservices architecture.
• Location: Based in the US (East Coast time zone slightly preferred).

Engagement Highlights

• Enterprise Production Scale: Move beyond proof-of-concept builds to deploy actual, living agent systems that automate mission-critical enterprise processes.
• Frontier Tech Stack: Work directly with state-of-the-art agent SDKs, Model Context Protocol, and multi-model ecosystems across Anthropic, OpenAI, and GCP.
• Remote Autonomy: Enjoy a fully remote US-based role with high execution agency and low friction.