Model Online / On-Site
Duration
Modules 5 modules

Course Objective & Scope

Traditional OCR and template rules regularly break on multi-column PDFs, merged-cell tables, stamped or signed pages and scanned contracts. This course teaches hybrid document intelligence pipelines that combine Vision-Language Models (VLMs) with modern layout parsers such as Docling, Azure Document Intelligence and LlamaParse.

Participants extract structured data from enterprise documents via JSON Schema, ground each extracted field to page and bounding-box coordinates, write cross-field validation rules, and build a document analysis application that routes low-confidence fields to human review.

Who Should Attend?

  • Developers automating document-heavy workflows in finance, legal, insurance, logistics or HR
  • Data engineers, AI engineers and RPA/automation specialists
  • Technical analysts building document processing products
  • Prerequisite: basic Python and JSON/API familiarity

Key Takeaways

  • Combine traditional OCR, layout parsers and direct VLM extraction effectively
  • Extract structured data from complex tables, charts, stamps/signatures and handwriting
  • Verify extracted values with deterministic business rules and cross-totals
  • Ground every extracted field to source page coordinates for auditability
  • Establish a field-level accuracy (F1/exact match) and latency measurement plan

Curriculum

Document Intelligence Architectures: From OCR to VLMs
  • Comparing Docling, Marker, Azure Document Intelligence and direct VLMs
  • Document classification, page splitting and orientation correction
Complex Layouts, Tables and Charts
  • Multi-page tables and nested headers
  • Detecting diagrams, technical drawings, stamps, signatures and checkboxes
Schema-Driven Extraction and Visual Grounding
  • Defining required fields, types and formats with JSON Schema
  • Linking extracted values to page numbers and bounding boxes
Validation, Confidence Scoring and Human Review
  • Subtotal/tax reconciliation, date logic and master-data lookups
  • Routing uncertain fields to side-by-side operator verification
Hands-on Lab and Quality Measurement
  • End-to-end extraction pipeline on representative enterprise documents
  • Field-level accuracy, cost and per-page latency benchmarking

Other Courses in This Category

  • Multimodal RAG: Indexing and Retrieval Across Text, Tables and Visuals

    Multimodal RAG architecture over enterprise documents containing tables, charts, diagrams and images: visual-text indexing, ColPali and image embeddings, hybrid retrieval and visually grounded answer generation.

    Multimodal & Conversational AI Corporate
    Live Course
    5 modules
    View
  • Conversation Design for AI Assistants and Voice Agents

    Conversation design for LLM-based chat and voice assistants: persona and tone guidelines, multi-turn dialogue flows, disambiguation, repair strategies and seamless handoff to human agents.

    Multimodal & Conversational AI Corporate
    Live Course
    5 modules
    View
  • AI Customer Service Agents and Contact Centre Transformation

    Designing AI customer service agents and human-agent assist systems powered by knowledge bases (RAG), CRM/order tools and policy guardrails, from contact centre architecture to production rollout.

    Multimodal & Conversational AI Corporate
    Live Course
    5 modules
    View
  • Voice AI Systems: STT, TTS and Conversational Audio Pipelines

    Speech-to-Text (STT), LLM and Text-to-Speech (TTS) pipelines for enterprise voice AI: low-latency audio processing with Whisper, Deepgram, ElevenLabs and Cartesia, acoustic challenges and telephony integration.

    Multimodal & Conversational AI Corporate
    Live Course
    5 modules
    View
  • Building Real-Time Multimodal and Voice Assistants

    Real-time AI assistants that interact instantaneously over voice, camera and screen sharing: full-duplex streaming, barge-in handling, live visual context, in-flight tool calling and resilient session management.

    Multimodal & Conversational AI Corporate
    Live Course
    5 modules
    View
  • Computer Use Systems: Screen-Based AI Agent Architecture

    Computer-use systems that operate desktop and browser environments via screenshot perception, coordinate grounding and keyboard/mouse actions: Claude Computer Use, OpenAI CUA, sandboxed VMs and safe execution.

    Multimodal & Conversational AI Corporate
    Live Course
    5 modules
    View

Get in touch