Model Online / On-Site
Duration
Modules 6 modules

Course Objective & Scope

The single most common cause of poor RAG quality is not the LLM or vector database, but mangled PDF tables, split clauses and context-stripped chunks created during ingestion. This course covers layout-aware extraction from messy enterprise formats (multi-column PDFs, scanned files, spreadsheets, HTML, Confluence/SharePoint) and advanced chunking architectures.

Participants test Docling, Marker/Unstructured and VLM-based parsers on challenging documents from their own organisation, building an ingestion pipeline with structural, parent-child and contextual chunking.

Who Should Attend?

  • Data and AI engineers building ingestion pipelines for RAG and enterprise knowledge bases
  • Software developers processing complex regulations, contracts, technical manuals and financial reports
  • Data scientists improving retrieval accuracy through better data preparation
  • Prerequisite: Python data processing experience and basic familiarity with RAG

Key Takeaways

  • Select the right parser for multi-column PDFs, complex tables and scanned documents
  • Index tables and figures without losing structure by combining Markdown/JSON with natural-language summaries
  • Apply heading-aware, parent-child or contextual chunking matched to document structure
  • Strip header/footer noise, deduplicate content and enrich chunks with hierarchical metadata
  • Design incremental ingestion pipelines that update and purge stale chunks via content hashes

Curriculum

Document Parsing and Layout Analysis
  • Why naive PDF readers fail: reading order, multi-column layouts and footnotes
  • Comparing Docling, Unstructured, Surya/Marker and VLM-based document parsers
Indexing Tables, Lists and Visual Elements
  • Preserving table integrity: HTML/Markdown representations, header propagation and summary embeddings
  • Indexing diagrams, charts and embedded figures via VLM captions
Chunking Strategies: Fixed, Structural and Semantic
  • Token-based, recursive and Markdown/HTML heading-hierarchy chunking
  • Embedding-breakpoint semantic chunking and when it actually helps
Preserving Context: Parent-Child, Late Chunking and Contextual Retrieval
  • Small-to-big retrieval: matching on fine chunks while passing parent sections to the LLM
  • Contextual Retrieval (prepending document-level context per chunk) and late chunking
Cleaning, Deduplication and Incremental Ingestion Pipelines
  • Boilerplate header/footer removal, near-duplicate detection (MinHash/SimHash) and PII scrubbing
  • Content-hash tracking for incremental updates and orphan chunk cleanup
Lab: Ingestion Pipeline on Complex Corporate Documents
  • Hands-on parsing and chunking experiments on internal PDFs and tables
  • Measuring how chunking choices impact downstream retrieval recall

Other Courses in This Category

  • Citations, Grounding and Hallucination Control in RAG Systems

    Engineering RAG systems for strict grounding, verifiable inline citations, conflict resolution across sources, calibrated abstention when evidence is missing and automated faithfulness verification.

    RAG & Enterprise Knowledge Systems Corporate
    Live Course
    6 modules
    View
  • Advanced RAG Advanced RAG

    Enterprise RAG architectures covering BM25-dense hybrid retrieval, ColBERT, neural reranking, self-correcting Agentic RAG, GraphRAG, Text-to-SQL, and GPU-accelerated serving.

    RAG & Enterprise Knowledge Systems Corporate, Individual
    Live Course
    10 modules
    View
  • Hybrid Search and Reranking for High-Precision RAG

    Boosting RAG retrieval precision and recall by combining BM25/SPLADE lexical search with dense vectors via RRF, query rewriting (multi-query, HyDE) and cross-encoder or ColBERT reranking.

    RAG & Enterprise Knowledge Systems Corporate
    Live Course
    6 modules
    View
  • Metadata Architecture and Permission-Aware Retrieval in RAG

    Designing metadata schemas, dynamic filter extraction and RBAC/ABAC/ReBAC security-trimmed retrieval so enterprise RAG assistants never leak unauthorized documents.

    RAG & Enterprise Knowledge Systems Corporate
    Live Course
    6 modules
    View
  • Agentic RAG: Multi-Step, Corrective and Adaptive Retrieval Architectures

    Building Agentic RAG systems with LangGraph that plan sub-queries, route across heterogeneous data sources, grade retrieved evidence, self-correct queries (CRAG/Self-RAG) and bound latency.

    RAG & Enterprise Knowledge Systems Corporate
    Live Course
    6 modules
    View
  • GraphRAG and Knowledge-Graph-Enhanced Retrieval Systems

    Combining knowledge graphs (Neo4j, property graphs), entity-relationship extraction, community summaries (Microsoft GraphRAG) and vector search to answer multi-hop and corpus-wide enterprise queries.

    RAG & Enterprise Knowledge Systems Corporate
    Live Course
    6 modules
    View

Get in touch