AI & ML Staff Engineer

About the role

dLocal is a fintech company providing payment infrastructure for global merchants in emerging markets. It is looking for an AI & ML Staff Engineer to act as the senior technical reference for building, operating and evolving its ML and AI infrastructure. The role enables Data Science and AI teams to move models and AI services from idea to production reliably, observably and compliantly, and owns the technical direction of the MLOps stack.

Key responsibilities

• Define and evolve the end-to-end ML platform architecture (data, training, registry, serving, monitoring, governance) used by multiple squads.
• Design standard patterns for reproducible training pipelines, experiment tracking, model packaging, versioning and promotion flows, and safe online/batch inference rollouts (canary, shadow, rollback).
• Balance reliability, performance and cost for ML workloads with SRE/Infra and Finance/FinOps.
• Act as the go-to expert for complex MLOps questions and review designs and deployments against platform standards.
• Partner with Fraud, Anomaly and other squads on SLAs/SLOs, logging, metrics and alerts, and contribute to on-call readiness.
• Use agents and AI services to automate MLOps workflows such as feature platform operations, training/eval pipelines, alert triage and cost optimization.
• Set technical standards for model/data access control, PII handling, redaction, environment separation and auditability, and align with InfoSec and Architecture.
• Mentor MLOps and Data/ML engineers and lead design and architecture reviews.

Required profile

• Solid experience owning or designing MLOps platforms or ML infrastructure used by multiple teams.
• Strong background in distributed systems and data/stream processing.
• Ability to communicate with technical and non-technical stakeholders.
• Plus: prior role as Technical Referent, Tech Lead or Architect, and experience rolling out AI assistants in engineering organizations.

Required skills

• Production ML pipelines, model registries, CI/CD for models and data, and online/batch inference at scale.
• Databricks, including Feature Store (ideally Unity Catalog), MLflow and batch/real-time deployment.
• Cloud ML platforms (e.g. SageMaker, Vertex AI) and container-based deployments.
• Observability: metrics, logs, traces, data/model drift and quality checks.
• Plus: LLM/AI infrastructure (gateways, vector stores, evaluation harnesses).