AI Engineer – Agents

Omie's purpose is to bring prosperity to the Brazilian entrepreneurial ecosystem. Therefore, we look for people with a growth mindset, valuing profiles that have adaptability, a collaborative sense, and are motivated by innovation and continuous learning. Be part of the history of one of the fastest growing tech companies in the country! 🚀 Join Omie. Develop, evolve, and maintain AI agents with LangChain and LangGraph: state and flows, tool calling, memory, human-in-the-loop, and multi-agent architectures. Integrate LLMs and embedding models via Amazon Bedrock, choosing the model and strategy based on quality, latency, and cost. Build RAG pipelines: document ingestion and chunking, embeddings, indexing, and search in Amazon OpenSearch Service and Amazon S3 Vectors (directly or via Amazon Bedrock Knowledge Bases), and self-hosted vector databases on EC2/EKS. Develop the APIs and services that expose the agents and connect them to the Omie ERP, databases, and legacy systems. Deploy and operate solutions on AWS (Lambda, Step Functions, EC2, EKS, and S3) using Infrastructure as Code and CI/CD. Create evaluations (evals) and implement observability: test datasets, quality metrics, tracing, and cost and latency monitoring in production. Apply guardrails and security/responsible AI practices: prompt injection mitigation, control of tools and data each agent can access, and personal data protection (LGPD). Work