AI Data Engineer III

About Rimini Street, Inc .

Rimini Street, Inc . (Nasdaq: RMNI), a Russell 2000® Company, is a global provider of end-to-end enterprise software support, products and services, the leading third-party support provider for Oracle and SAP software and a Salesforce and AWS partner. The Company has operations globally and offers a comprehensive family of unified solutions to run, manage, support, customize, configure, connect, protect, monitor, and optimize enterprise application, database, and technology software. To date, over 5,300 Fortune 500, Fortune Global 100, midmarket, public sector, and other organizations from a broad range of industries have relied on Rimini Street as their trusted enterprise software solutions provider.

We are actively seeking an AI Data Engineer III . This is a remote role based in Brazil.
Position Summary

The AI Data Engineer is responsible for building the knowledge layer of Rimini Street’s Agentic ERP Platform—the data pipelines, RAG (Retrieval-Augmented Generation) systems, and embedding infrastructure that give AI agents access to the right information at the right time. This role owns how knowledge is ingested, processed, indexed, and retrieved to support intelligent agent behavior.

Reporting to the Sr. Director, Engineering, this engineer designs the data architecture that powers agent intelligence—from extracting knowledge from Rimini Street’s 15+ years of support case history to building real-time retrieval systems for customer-specific context. The ideal candidate combines strong data engineering fundamentals with modern AI/ML knowledge, particularly in embeddings, vector search, and retrieval optimization.
Essential Duties & Responsibilities
RAG Pipeline Development

-
Design and build RAG pipelines that retrieve relevant context from knowledge bases to augment AI agent responses.

-
Implement chunking strategies optimized for different content types: support tickets, documentation, policies, transaction records, and email threads.

-
Develop hybrid retrieval approaches combining dense embeddings, sparse search (BM25), and metadata filtering.

-
Build query understanding and reformulation logic to improve retrieval relevance.

-
Implement retrieval evaluation frameworks to measure and optimize precision, recall, and relevance.

-
Design reranking pipelines that prioritize the most relevant results for agent consumption.

Embedding & Vector Infrastructure

-
Implement and manage vector storage using PostgreSQL with pgvector extension, including index optimization for search performance.

-
Evaluate and select embedding models appropriate for enterprise content (technical documentation, business processes, ERP terminology).

-
Build embedding pipelines that process documents at scale with appropriate batching and error handling.

-
Implement incremental indexing strategies for real-time updates without full reprocessing.

-
Design multi-tenant vector architectures that isolate customer data while enabling efficient search.

-
Monitor and optimize vector search performance: latency, accuracy, and resource utilization.

Data Ingestion & Processing

-
Build data pipelines to ingest knowledge from diverse sources: Salesforce support tickets, ServiceNow cases, documentation repositories, email archives, and ERP transaction logs.

-
Implement ETL processes that clean, normalize, and enrich raw data for AI consumption.

-
Develop document processing pipelines: PDF extraction, HTML parsing, structured data normalization.

-
Build connectors to source systems including Salesforce, ServiceNow, SharePoint, and Confluence.

-
Implement data quality monitoring and alerting for ingestion pipelines.

-
Design data lineage tracking to understand how knowledge flows from source to agent consumption.

Knowledge Architecture

-
Design the knowledge architecture that organizes information across the Four-Spoke model: Policy Intelligence, Institutional Memory, Rimini Collective Intel