Document Chunker

Software componentKnowledge & DataKnowledge & DataVariation point (abstract)arc:DocumentChunker

A software component that splits raw documents into chunks for embedding and entity extraction.

Responsibility. Splits raw documents into processable chunks.

Direct neighbourhood (hover for relationship types)

Variants

Variant	When to choose
Fixed-Length Chunker	Choose only when content has no usable structure or timing data; the chapter presents it as the naive baseline that fractures semantic units.
Hierarchical Chunker	Choose when document structure (sections, subsections) must be preserved so retrieved chunks keep their structural context, accepting more complex retrieval logic.
Overlapping Window Chunker	Choose when important context spans chunk or section boundaries and continuity must be preserved, accepting multiplied storage.
Semantic Boundary Chunker	Choose for text documents with structure, splitting at section headers and paragraph breaks rather than arbitrary token counts.
Time-Indexed Transcript Chunker	Choose for audio transcripts lacking paragraph breaks or headers, when retrieved segments must link back to exact moments in the recording.
Topic-Shift Chunker	Choose when chunk coherence matters more than uniform chunk size.

Relationships

receives data from dynamic

sends data to dynamic

Chunk Metadata Extractor Ch6.3A
Embedding Service abstract Ch1.7A Ch3.10 +1
Entity Recognizer Ch1.7A
Text Embedding Service Ch6.4

is orchestrated by control

Ingestion Pipeline Orchestrator Ch2.7 Ch6.2B +4

Design guidance

SHOULD produce 300-500 token segments rather than retrieving entire documents.
SHOULD NOT split on fixed sizes alone; chunks beginning mid-context lose referents and degrade retrieval.
SHOULD choose chunk size, overlap and boundary logic from document characteristics and retrieval requirements.
SHOULD count tokens with the embedding model's tokenizer rather than character approximations, especially for code and non-English text.
SHOULD discard trailing fragments below a minimum chunk size.

Classification

Patterns: Boundary seeking (paragraph, then sentence, then word)Minimum chunk size enforcementToken-accurate counting
Risks mitigated: Context loss from undersized chunksGeneric, unspecific embeddings from oversized chunks

Sources

Ch1.7A: T. Nguyen, "Relational Reasoning with Knowledge Graphs - The Fundamentals, Integration, and Extraction," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 1.7A. ISBN: 9798244538229.
Ch1.7B: T. Nguyen, "Relational Reasoning with Knowledge Graphs - Hybrid RAG+KG Integration," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 1.7B. ISBN: 9798244538229.
Ch2.7: T. Nguyen, "Multimodal RAG Approaches," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 2.7. ISBN: 9798244538229.
Ch3.10: T. Nguyen, "Efficiency Metrics," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 3.10. ISBN: 9798244538229.
Ch5.8: T. Nguyen, "Semantic Memory," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.8. ISBN: 9798244538229.
Ch6.2B: T. Nguyen, "Production Vector Database Deployment," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 6.2B. ISBN: 9798244538229.
Ch6.3A: T. Nguyen, "ETL Pipeline Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 6.3A. ISBN: 9798244538229.
Ch6.4: T. Nguyen, "Data Quality Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 6.4. ISBN: 9798244538229.
Ch6.5: T. Nguyen, "Production RAG Systems," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 6.5. ISBN: 9798244538229.
Ref2.07: NVIDIA Developer, "Building multimodal AI RAG with LlamaIndex, NVIDIA NIM, and Milvus | LLM app development," YouTube. Accessed: Sep. 26, 2026. [Online Video]. Available: https://www.youtube.com/watch?v=NaT5Eo97_I0