Build Automated SharePoint → Azure AI Search Pipeline with OCR and Document Retrieval

We are seeking an experienced Azure / Microsoft 365 developer to build an automated pipeline that connects our SharePoint document repositories with Azure AI Search and Azure OpenAI. The system will automatically ingest and process documents, then retrieve content from specific SharePoint folders and combine it with a curated “authorities” folder for drafting structured outputs. A workflow diagram has been attached to illustrate the high-level system architecture. Sample files (both text-based and scanned PDFs) can be provided for testing. ________________________________________ Scope of Work • Connect to Microsoft 365 SharePoint via Microsoft Graph API to ingest files. • Automate ingestion with Graph webhooks + Azure Functions (delta sync only re-indexes changed/new files). • Implement OCR automation for scanned PDFs using Adobe PDF Services API (we already have credentials) or Azure Form Recognizer. • Store embeddings and text chunks in Azure AI Search, with separate namespaces for: o Document folders (e.g., “Project A” folder) o Authorities/reference library (rules, standards, references) • Enforce retrieval logic: o Facts/content come only from project document folders. o Rules/references come only from the authorities folder. • Quarantine workflow: flag low-confidence OCR → move to “Review” folder in SharePoint → send Teams notification to staff. • Integrate with ChatGPT Team Plan to draft outputs (e.g., reports, summaries, motions, structured documents) with citations linked back to SharePoint. • Ensure security and compliance: mirror SharePoint ACLs (on-behalf-of tokens), encrypt vectors at rest, maintain audit logs. ________________________________________ Required Skills • Strong experience with Microsoft Graph API (SharePoint/OneDrive integration). • Hands-on with Azure AI Search (vector indexes, filters, namespaces). • Knowledge of Azure Functions and event-driven automation. • OCR integration: Adobe PDF Services API or Azure Form Recognizer. • Familiarity with Azure AD authentication/permissions. • Bonus: experience with document-heavy or compliance-driven workflows. ________________________________________ Deliverables • Working ingestion pipeline (pilot with one folder + one authorities library). • Automation layer (OCR + delta sync + Teams alerts). • Documentation for ongoing maintenance (how staff add new folders, handle quarantine files, review logs). • ***All code delivered via GitHub repo (or zip) with a clear README, infrastructure-as-code (ARM template or Terraform), and deployment instructions.*** ________________________________________ Environment • Microsoft 365 Business Premium • SharePoint + Teams • Adobe Acrobat Pro (with PDF Services API credentials) • ChatGPT Team Plan ________________________________________ What We’re Looking For • A developer who can deliver a robust, production-ready system, not just a prototype. • Clear communication, documentation, and ability to hand off to non-technical staff. ________________________________________ Screening Questions (Required) • Have you implemented Microsoft Graph API webhooks with SharePoint before? Please provide an example. • Have you built an ingestion pipeline into Azure Cognitive Search / Azure AI Search? • What OCR solution(s) have you integrated before (Adobe PDF Services API, Azure Form Recognizer, or other)? • Please include links to previous Azure AI Search or SharePoint Graph API pipeline projects in your bid.

Project ID: 39730016