Senior - AI Engineer

Architect and build the software solution that meets the client's demand, in any stack, from architecture to tested code, using the agentic SDLC as the way of working and AI as the main build lever, and remaining accountable for the application delivered in production. Builds and orchestrates their own agents, skills, subagents, commands and harness that accelerate building the application and reuses them across the team, instead of consuming assistants passively. Answers for the quality and the delivery of what they produce, applying critical judgment to AI-generated outputs and pragmatic technical design to avoid overly complex solutions, grounded in modern software architectures, instrumenting observability at system level and using applied AI as a supporting repertoire, focusing on fine-tuning models rather than training them from scratch. When the client's solution includes an AI component, consumes foundation models through APIs and assembles the solution with RAG, tool orchestration, evals and multi-agent designs, structuring the engineering and configuring the context data pipelines and the orchestration and runtime frameworks the agents run on. Uses evals to assess both the probabilistic behavior of the solution and the agents built to accelerate development.

Responsibilities:

- Architect and build the software solution that meets the client's demand, in any stack, from architecture to tested code, using the agentic SDLC as the way of working and AI as the main build lever, and remaining accountable for the application delivered in production.

- Build, customize and orchestrate their own agents, skills, subagents, commands and harness that accelerate building the application, reusing them across the team instead of consuming assistants passively, driving the continuous modernization of software development processes and practices, and reducing the time between building and delivering (lead time).

- Master and apply modern software architecture and engineering fundamentals (microservices, event-driven systems, queues, CI/CD, TDD and automation) to build, assess and critique the code and the solutions generated.

- Translate functional requirements into a viable, simple and scalable technical solution, making intentional choices instead of passively accepting AI suggestions, to avoid over-engineering.

- Design and assess, with evals, tests and red teaming (hallucinations, injections and leaks), both the solution and the agents and skills they build, in order to validate the probabilistic behavior of the solution before delivery.

- Make token economy decisions in the solution design (context size, RAG strategy and number of agent calls), to balance quality and inference cost.

- When the solution includes an AI component, architect and build the agentic solution: agents through APIs with RAG, tool orchestration, orchestration and runtime frameworks, and system prompts and guardrails in the application.

- When the solution includes an AI component, manage the engineering and the context data pipelines, structuring the integration of internal sources and vector databases to feed agentic systems with reliable, high quality inputs.

- Implement observability instrumentation within the application and agent code, continuously monitoring system behavior to act on anomalies and degradation.

- Review and critique AI-generated outputs, ensuring adherence to technical, quality and business requirements before delivery.

- Delegate specialized tasks to the agents they build, working with a generalist view across the whole development line, while sustaining technical depth and authority in their specialty stack to critically assess and validate outputs.

- Deliver the solution in a production-ready state, aligned with the AI Deployment Engineer on deployment and operation requirements.

- Apply ethics, privacy, data security and compliance practices, to keep the solutions within the client's and CI&T's rules.

- Define and lead the solution design of complex problems, raising the technical quality of the team.

- Own the solution end to end in complex scenarios (multiple systems, integrations and models), from architecture to delivery in production.

- Establish and reuse across the team (not only for themselves) the agents, skills and harness.

- Structure orchestration and runtime frameworks, ensuring reliable inputs and behavior at scale.

- Guide prompt engineering, evals and red teaming decisions at scale.

- Act as a technical reference and mentor to peers, reviewing code and standards.

- Collaborate with the AI Deployment Engineer and the AI Orchestrator, making building blocks and infrastructure available and ready.