AI Specialist (Backend / Vertex AI)

About the role

We are looking for a Backend Developer with a focus on Artificial Intelligence to train, evaluate, and put into production the AI models of our intelligent platform for managing field operations.

You will work on a large-scale project in the public education system. There, our AI agents calculate indicators from inspections, automatically triage tickets, and answer, in natural language, questions from managers about operational data. Your mission is to ensure that these responses are correct, safe, and useful for thousands of users.

Why this role

• • AI in production, not in the lab: the models you train will be used every day, at scale.
  • Full cycle: you go from training data to deployment, with technical autonomy and direct contact with the CTO.
  • Current stack: Google Cloud, Vertex AI, and Gemini models.
  • Real impact: better decisions about public school infrastructure.

Responsibilities

• • Train, tune, and version the AI models embedded in the platform, such as those for calculating indices from inspections and automatically triaging tickets.
  • Develop and evolve the conversational agent that answers natural language queries about platform data.
  • Build and maintain test sets with expected inputs and outputs, evaluate model responses, identify errors and hallucinations, and adjust prompts and models.
  • Deploy models to scalable endpoints and monitor quality, latency, token consumption, and cost.
  • Develop APIs in Node.js and Python that integrate AI models with the platform.
  • Implement protections (guardrails), auditable logs, and human validation workflows for automated decisions.
  • Turn user feedback into continuous improvements to the models, together with product and operations teams.

Requirements

• • Proficiency with the model lifecycle in Vertex AI, including training, versioning in Model Registry, deployment to scalable endpoints, and use of Vertex AI Studio.
  • Experience with language models (Gemini or similar) in production: prompt engineering, function calling, and responses based on application data (RAG).
  • Experience evaluating AI models: test sets, quality metrics, and regression testing of prompts and models.
  • Experience with model fine-tuning.
  • Proficiency with Python for data and AI pipelines, and experience with Node.js in developing RESTful APIs.
  • Proficiency with PostgreSQL, with a focus on performance.
  • Hands-on experience with Google Cloud: Cloud Run, Cloud Functions, Cloud Storage, IAM, Cloud Logging, and Cloud Monitoring.
  • Experience with Docker, automated testing, code review, and CI/CD.
  • Attention to AI cost control, such as token consumption and endpoint usage.

Nice to have

• • Vector search and embeddings (Vertex AI Vector Search or pgvector).
  • Multimodal models, such as image analysis.
  • Security in AI applications, such as protection against prompt injection and anonymization of personal data (LGPD).
  • Kubernetes (GKE), Terraform, and messaging (Pub/Sub).
  • BigQuery.
  • Completed degree in Computer Science, Software Engineering, Information Systems, or related fields.