Jean-Dominique Stepek

Collection Curator

• Shaped the reporting workflow for an operator-facing analytics portal — filters, drill-downs, real-time collaboration via Legend State sync engine, and follow-up actions in one surface — replacing scattered tools and ad hoc data pulls.
• Built a configurable table interface for seasonal collection curation with custom column types (multiselects, dates, tags), enabling analysts to filter and organize product data without ad hoc spreadsheet work.
• Led frontend development of a PowerPoint-style presentation tool with React and Styled Components, including drag-and-drop, keyboard navigation, and export-to-PDF features.
• Created an internal GenAI chatbot using AWS Bedrock and RAG retrieval for answering platform-related user queries — cutting support tickets by 50%.
• Architected the production API backend pairing Express/Apollo with a reverse-proxied FastAPI service for AI workloads, using Prisma, Redis pub/sub, and Cognito-backed auth.
• Wrote MCP tool definitions for the analytics platform and built a demo surface showing user-driven vs agent-driven actions for stakeholder evaluation.

Supply Chain Forecasting App

• Spearheaded full-stack development of a predictive analytics platform with React, TypeScript, GraphQL, and Snowflake, reducing inventory waste by $30M/year.
• Designed the forecasting interface around fiscal season targets and receipt predictions, enabling inventory planners to adjust forecasts without falling back to spreadsheet workflows.

Sample Tracking

• Built the operational app for sample shipment tracking, status management, exports, and recurring reporting — consolidating scattered update threads into one surface for 16 stakeholders.
• Saved ~800 hours/year by replacing manual reconciliation of exports, shipment context, and recurring updates across separate tools.

• Co-founded and built a social analytics platform tracking community engagement patterns across social media APIs, with unified customer profiles and proprietary metrics.
• Contributed to data ingestion pipelines from social media APIs to DGraph using Go and AWS Lambda, and owned infrastructure: Dockerized services, CI/CD with AWS CodePipeline, EC2 and Lambda compute.