[Job-31203] AI Engineer Master, Brazil

We are building the client's AWS enterprise cloud foundation from the ground up, and we are building it with an Agentic Software Development Life Cycle (Agentic-SDLC). We are not looking for an architect who reviews what agents produced. We are looking for one who builds the agents, defines the intents they execute, and treats the delivery system itself as part of the architecture they own.

This is the senior technical seat on the account. You will own the architecture across workstreams — the multi-account AWS landing zone (Network, Shared Services, and Workload Non-Prod and Prod), the application architecture running on top of it, and the platform architecture that comes next. You will also lead the operational team as tech lead: setting the technical direction, unblocking engineers, and shipping code yourself. This is a hands-on role. The person who owns the diagram is the person who writes the CDK.

Responsibilities

- Architecture Ownership: Own the AWS enterprise architecture end to end — multi-account governance and guardrails, Transit Gateway topology, hybrid connectivity to on-premises data centers, centralized traffic inspection, hybrid DNS, and the security posture that holds it all together. You write the ADRs, you defend them to the client, and you revisit them when they stop being true.

- Tech Leadership with Your Hands On: Lead the operational team day to day — technical direction, design review, code review, and the architectural decisions that unblock delivery. You are also expected to build: CDK stacks, Lambda functions, agent tooling.

- Agent Development & Orchestration: Design and extend the agents that carry our delivery — intent decomposition, design artifact generation, implementation scaffolding, validation, and gate reporting. Author the capability instructions they run on so agents produce architecture-compliant work without a human restating the standards each time. When you see the same correction made twice, you encode it once.

- Intent-Driven Execution: Work within our Intent-Driven Framework, where intent describes client outcomes rather than task lists, ambiguity stops execution instead of getting interpreted, and every user-visible change ships behind a feature flag. High-velocity delivery in fixed-scope cycles, without scope creep or accumulated technical debt.

- Legacy Modernization: Help systematically retire a live VB6 desktop application by moving its functionality into cloud services, while the legacy system keeps running untouched. Coexistence design — read-only replication, parity proof, staged cutover — is a first-class deliverable, not an afterthought.

- Platform Direction: Lead the platform evaluation for the external digital front: understand what the client's dealers and commercial customers actually need, assess the options, and produce a recommendation leadership and the client can act on. Then design the reuse path — external identity, tenancy, API contracts that hold once a partner integrates against them.

- Client-Facing Architecture: Be the person in the room with the client's IT and security leadership explaining why the design is what it is, and the person who translates their constraints back into design changes.

Required Skills and Qualifications

- Cloud Architecture Depth: Extensive years in software engineering with architecture, on production systems you were accountable for. Serverless application architecture — API Gateway, Lambda, CloudFront, Cognito with SAML federation, Secrets Manager, CloudWatch and X-Ray.

- Infrastructure as Code, Exclusively: AWS CDK in TypeScript, with opinions about stack boundaries and construct design. No console provisioning, no drift.

- Software Engineering: Java and TypeScript at a level where you can review and write, not just discuss. Our stack is Java 25 with Spring Cloud Function on Lambda, React 19 with TypeScript, CDK in TypeScript. REST contract design, versioning, and backward compatibility. GitHub Actions with OIDC-based cloud authentication.

- AI/ML Proficiency: Hands-on experience with Prompt Engineering, RAG, and Agentic Frameworks (e.g. Claude Agent SDK, Strands, LangChain, or CrewAI). Tool and MCP integration, context engineering, and a real point of view on how agent output gets verified at a scale where you cannot read all of it. AWS Bedrock and AgentCore are a plus.

- Hybrid and Legacy Reality: Experience shipping something that had to coexist with a system nobody was allowed to touch.

Soft Skills

- Adaptive and Agentic Mindset: Ability to pivot quickly. You should be comfortable with a "startup within a corporate" feel where priorities shift based on impact, and where the highest-value thing this week was not on the plan two weeks ago.

- Architectural Thinking: Understanding of the "Agentic SDLC" and the ability to build systems that reduce human intervention over time. You should find it satisfying, not threatening, when something you used to do by hand becomes automatic.