[Job - 30900] Master Data Developer (Amazon Neptune), Colombia

Job description

At CI&T, we help large enterprises transform the potential of AI into real business impact with AI Deployment, AI-native execution, and tech-integrated business solutions. With 30 years of experience in technological transformation, we accelerate innovation with expertise in Agentic SDLC, Application modernization, Data & AI, Martech and Business strategy. We are 8,000 CI&Ters across more than 25 countries, collaborating to build solutions with real impact. AI is already part of how we work, evolve, and innovate every day. As CI&T grows its Data & Analytics Center of Excellence, we seek a talented and experienced Graph Database Developer to join a specialized team building an AI-powered fashion advisory platform for a leading client in the fashion and retail industry. This role is central to turning unstructured editorial content into a structured knowledge graph that powers real-time, intelligent style recommendations for end users attending formal events. The Graph Database Developer will work in a staff augmentation model, integrating with a cross-functional team that includes a Machine Learning Developer and a GenAI Agent Developer. This is a hands-on technical position requiring strong ownership of the graph data layer — from schema design through ingestion, entity resolution, and query performance — with direct impact on what the AI agent recommends to users. **Responsibilities:** - **Schema Design**: Define node labels and edge types that represent the relationships between designers, garments, style attributes, trends, occasions, and editorial articles. - **Entity Ingestion**: Build workflows that consume structured JSON payloads from the upstream data pipeline and load them into the graph database in batch mode. - **Probabilistic Entity Matching**: Match extracted item mentions (for example, a bag described by type, color, and material) to the correct product node in a catalog of 50,000+ SKUs by scoring attribute overlap, assigning a match confidence, and creating the relationship only when it exceeds a defined threshold. Deduplicate items that appear inconsistently across different articles. - **Editorial Signal Weighting**: Assign weights to relationships based on how prominently an item or trend was featured, distinguishing a dedicated feature from a passing mention. These weights determine what the AI agent surfaces first in its recommendations. - **Query Optimization**: Design and tune openCypher traversal queries, including multi-hop queries, to return relevant results in under three seconds for downstream Amazon Bedrock Agents. - **Cross-Functional Collaboration**: Work alongside the Machine Learning Developer and GenAI Agent Developer to align the graph structure with the retrieval and recommendation needs of the AI agent. - **Technical Validation**: Support technical interviews and validation of new team members joining the graph database workstream, as needed. **Requirements:** - Advanced English proficiency (C1 or above), with autonomy to communicate directly with international stakeholders - Solid experience with the property graph data model - Experience writing and optimizing openCypher queries, including multi-hop traversals under latency constraints - Experience designing graph schemas, evaluating trade-offs between different modeling approaches - Experience building entity resolution logic, including fuzzy matching, scoring algorithms, and deduplication across data sources - Strong Python skills for data engineering, including batch processing of structured data payloads - Ability to work independently in a fast-paced, staff augmentation setting with an established client team **Nice to Have:** - Solid experience with Amazon Neptune - Prior experience in fashion, retail, or e-commerce catalogs - AWS certification in databases or data analytics - Experience integrating graph databases with generative AI agents (for example, Amazon Bedrock Agents) - Familiarity with recommendation systems or ranking/weighting logic for content surfacing #LI-JP3