Engenheiro(a) de IA - Aplicações Agênticas (Pleno)

Accountabilities:

• Develop specialized AI agents coordinated by supervisors, implementing agentic loops such as ReAct and maintaining auditable records of decisions and actions.
• Build MCP Servers that connect AI agents with operational APIs and machine learning models through well-defined OpenAPI contracts.
• Implement MCP Hosts using frameworks such as LangChain or Semantic Kernel and integrate LLM capabilities through Azure OpenAI.
• Design and implement Human-in-the-Loop workflows for decisions that require human validation, oversight, or intervention.
• Establish secure agentic architectures, including agent-specific identities, OAuth2 authentication, tool-level permissions, idempotency, prompt-injection protection, and safeguards against PII leakage.
• Apply PromptOps practices by versioning prompts and implementing agent observability through LangSmith or similar platforms.
• Integrate AI applications with event-driven architectures using technologies such as Azure Service Bus and Kafka, while ensuring reliable communication between services.

Requirements

• Bachelor’s or postgraduate education in Data Science, Computer Science, Statistics, Engineering, Mathematics, Economics, or a related field.
• Professional proficiency in English or Spanish, with the ability to communicate effectively in an international client environment.
• Strong programming skills in Python and/or C#/.NET, including experience with testing, version control, software contracts, and writing maintainable, readable code.
• Hands-on experience developing LLM-powered applications, including tool or function calling, agent orchestration, prompt engineering, and strategies for managing non-deterministic model behavior.
• Solid understanding of system integration and REST API development or consumption, including OpenAPI/Swagger, OAuth2 authentication, idempotency, webhooks, and callbacks.
• Experience enabling secure LLM access to data, including natural-language-to-query workflows with appropriate scope controls, validation, permissions, and auditability.
• Understanding of application security for LLM-based systems, including prompt injection, data leakage, and permission management at the tool level.
• Experience with MCP, LangChain or Semantic Kernel, Azure OpenAI, multi-agent architectures, Human-in-the-Loop patterns, agent observability, Kafka, Azure Service Bus, or .NET is considered a strong advantage.
• Familiarity with regulated environments, LGPD requirements, and auditing of automated decisions is also valuable, along with a proactive, analytical, collaborative mindset and a strong interest in emerging AI technologies.

Benefits

• PJ contractor model with a focus on building long-term professional partnerships.
• Fully remote work.
• 8-hour workday, totaling 168 hours per month.
• Bank of hours with quarterly settlement and payment of positive balances.
• Variable compensation after four months of contract.
• Flexible benefits card with R$700 per month for food allowance through Flash.
• Paid time off: 7 consecutive days after 1 year of contract, 15 days after 2 years, 21 days between 5 and 7 years, and 30 days from 8 years onward.
• Birthday month day off through the “Sexta Presente” benefit.
• Wellhub partnership.
• Open English partnership.
• FIAP partnership for professional development.
• Partnerships with clinics and psychology services.
• Opportunities to work with innovative AI technologies and challenging digital solutions in an international environment.