Data Engineer + Data Scientist for SAP Pharma Manufacturing AI PoC

TITLE • Build AI-Powered Manufacturing Quality Intelligence PoC – Data Engineering + Data Science + LLM ABOUT THE PROJECT We are a publicly listed AI company building an AI-powered Manufacturing Quality Intelligence & Process Excellence platform for the pharmaceutical industry. We need a senior freelancer (or a two-person team) to build a working Proof of Concept (PoC) that will be demonstrated to enterprise pharma decision-makers. The end client is confidential. IMPORTANT: This is NOT an SAP implementation, SAP HANA migration, ABAP, BASIS, or SAP functional consulting project. This is primarily a Data Engineering + Data Science + AI application build that uses SAP manufacturing and quality data as input. SAP knowledge is required only to understand the manufacturing data—not to implement SAP. WHAT YOU WILL BUILD A working PoC that analyzes one pharma manufacturing quality process—Batch Review & Release—and answers: • Where is time being lost? • What causes rework and delays? • What should be improved? Scope includes: 1. Data Pipeline - Ingest SAP-style CSV/Excel extracts - Generate realistic synthetic manufacturing event data - Build a clean event-log model 2. Process Reconstruction (Process Mining) - Rebuild actual process flows - Detect variants - Detect loops - Identify bottlenecks - Identify handoffs 3. KPI Analytics Engine Compute metrics including: - Batch Release Cycle Time - Waiting Time - Processing Time - Rework % - Right First Time % - Handoffs 4. Interactive Dashboards (~7 Screens) - Process Health Score - KPI Performance - Process Flow Visualization - Site/Product Drilldowns - Delay Analysis - Rework Analysis - Intervention Tracking 5. AI Copilot Build an LLM-powered assistant that: - Answers questions about the analytics - Uses tool/function calling - Never hallucinates numbers - Supports scenario analysis No live SAP integration is required. The PoC will work entirely from SAP extracts and synthetic data. RESPONSIBILITIES • Design the data model • Build the data pipeline • Generate synthetic data • Implement process mining logic • Build KPI computation engine • Develop dashboards • Build the AI Copilot • Deploy on a cloud test server • Provide daily updates • Deliver midpoint demo • Complete documentation and source-code handover REQUIRED SKILLS • Expert Python • Expert SQL • Data Engineering • Data Modeling • Process Mining / Process Intelligence • KPI Analytics • Dashboard Development (Streamlit / Dash / Power BI) • AI / LLM Integration • SAP Manufacturing & Quality Data understanding • Experience building complete AI analytics platforms STRONGLY PREFERRED • Pharmaceutical Manufacturing • Life Sciences • Batch Review & Release • QA/QC • GMP • Deviations / OOS • 21 CFR Part 11 awareness NICE TO HAVE • Databricks • Snowflake • Azure • Streamlit • Power BI • SAP BTP • SAP APIs • OData • MES • QMS • LIMS NOT REQUIRED Please DO NOT apply if your experience is primarily in: • ABAP • SAP BASIS • SAP Functional Consulting • SAP HANA Migration • SAP Implementation Programs TEAM STRUCTURE We are open to: Option 1 One senior freelancer covering both Data Engineering and AI/Data Science. Option 2 A two-person team: • Data Engineer • Data Scientist Please specify which model you are proposing. ENGAGEMENT • Fixed Price • Milestone Based • Fast-turnaround PoC • Midpoint demo mandatory • 100% IP ownership transfers to us • 1 week post-delivery bug-fix support • Strong possibility of a long-term engagement after successful PoC DELIVERABLES 1. Working end-to-end PoC 2. Data pipeline 3. Event-log data model 4. KPI Analytics Engine 5. Process Mining Engine 6. ~7 Interactive Dashboards 7. AI Copilot 8. Source Code 9. Documentation 10. Deployment Guide 11. Handover Session IDEAL CANDIDATE You have built complete AI-powered analytics platforms—from raw data pipelines to dashboards to conversational AI. Experience in Pharma, Manufacturing, or Life Sciences is highly preferred. You understand that KPIs must be computed deterministically and explained by the LLM—not generated by it. You can demonstrate similar work through a GitHub repository, portfolio, or live demo.