Data Engineer – Oracle to PostgreSQL Re-Platform

This is a fully remote position, open to applicants in Brazil.

📋 Description

• Oversee the complete delivery of a production LLM agent, managing everything from requirements to deployment while justifying architecture and trade-offs directly to the client.

• Develop and enhance agent orchestration using LangGraph, LangChain, or similar technologies, focusing on routing, tool invocation, planning, synthesis, and state management.

• Integrate tools over MCP and expand a growing tool surface, accommodating BM25, hybrid, or vector retrieval as scaling demands evolve.

• Operate models on AWS Bedrock and Bedrock AgentCore, which includes model selection/routing, implementing guardrails, memory management, and adhering to regional residency profiles.

• Create an evaluation harness featuring golden sets, utilizing LLM as a judge, and establishing quality gates that are integrated into CI processes.

• Instrument systems using OpenTelemetry for tracking per-session token usage, costs, and latency attribution.

• Optimize costs and reduce latency through model routing, prompt caching, payload pruning, and parallelizing independent calls behind a regression gate.

• Develop circuit breakers, fallbacks, and mechanisms for handling dead letters.

• Design and stage safe write-actions with minimal privilege permissions, human-in-the-loop approvals, plan versioning, audit trails, and rollback capabilities.

• Implement changes behind feature flags, initially releasing them to a limited group.

⛳️ Requirements

• A minimum of 5 years in software engineering, with at least 2 years in a truly senior position.

• Extensive experience in Python within production environments.

• Capability to learn TypeScript or Go as necessary.

• Practical hands-on experience in LLM application engineering, including prompt design, tool/function invocation, structured output management, context management, and token budgeting.

• Background in building or managing agent orchestration utilizing LangGraph, LangChain, or a custom framework that goes beyond single-prompt calls.

• Proven production experience managing an LLM/agent platform such as AWS Bedrock, Google Vertex AI, or Azure AI Foundry, including model invocation, streaming, guardrails, and agent tooling.

• Familiarity with evaluation harnesses, golden/reference sets, and vector or hybrid search in production, along with OpenTelemetry-based distributed tracing that includes token, cost, and latency attribution.

• Production-level experience with AWS, covering IAM, networking, storage, and observability.

• CI/CD pipelines managed as versioned code.

• Experience with Terraform in production settings.

• Proficient in English, with the ability to defend system design and technical choices directly during client calls.

🏝️ Benefits

• Competitive salary and performance-based bonuses.

• Opportunities for professional development and growth.

• Flexible work environment with remote work options.

• Comprehensive health and wellness benefits.

• Engaging company culture focused on innovation and teamwork.