AI Engineer
Headquarters: Remote URL: https://www.toptal.com/ We're looking for engineers to help build and productionize AI systems on AWS Bedrock AgentCore — real conversational AI, RAG pipelines, and agent architectures that go well beyond
Headquarters: Remote URL: https://www.toptal.com/
About the Role
We're looking for engineers to help build and productionize AI systems on AWS Bedrock AgentCore — real conversational AI, RAG pipelines, and agent architectures that go well beyond proof-of-concept and serve live traffic and real users. Whether your strength is on the AI application side (agents, RAG, orchestration) or the platform side (deployment, observability, security), this role sits at the center of turning working demos into production-grade, reliable systems on Bedrock's agentic stack.
If you've built and shipped on AgentCore specifically — not just Bedrock in general — and want your work to run in production rather than sit in a notebook, this is built for that.
What You'll Do
Design, build, and deploy conversational AI systems, chatbots, and AI agents using AWS Bedrock AgentCore Architect and ship production-grade RAG (Retrieval-Augmented Generation) systems — not prototypes, but systems serving live traffic Build and deploy LLM applications primarily on AWS Bedrock and AgentCore, with Azure OpenAI or equivalent platforms as secondary context Develop and orchestrate agent architectures within AgentCore, using frameworks such as LangChain, LangGraph, or LlamaIndex where applicable Build and maintain MCP (Model Context Protocol) server integrations to extend AgentCore agent capabilities Design and build the production service layer around AgentCore (Lambda, API Gateway, IAM, DynamoDB, OpenSearch, or equivalents) Establish CI/CD pipelines and manage development, beta, and production environments for AgentCore-based services Implement observability for AgentCore agents: tracing, dashboards, per-turn cost and latency metrics, error rates, and audit trails Implement key security controls — data-leakage protection, session isolation, auth/authz boundaries, secure prompt/response storage Write clean, maintainable, production-quality Python across the AI application and platform stack Monitor, evaluate, and iterate on agent, RAG, and platform performance in production Stay current with fast-moving developments in Bedrock, AgentCore, and agentic AI systems, and bring relevant advances into the project
What You Bring
Posted July 27, 2026