DakkaBot

As explained in Machine Learning Platform Engineering

DakkaBot is an internal documentation assistant for querying company documentation with natural language.

DakkaBot is an internal documentation assistant that lets engineers query company knowledge in natural language. It applies Retrieval-Augmented Generation (RAG) to enterprise documentation, extending established MLOps practices for generative AI. section 1.4.2 section 12.0

A production DakkaBot workflow combines:

• Document ingestion for technical guides, API documentation, runbooks, and architectural decision records. section 1.4.2 section 12.2
• Vector embeddings and semantic search for finding relevant documentation before generation. section 1.4.2 section 12.0
• Prompt engineering, safety guardrails, cost optimization, and semantic evaluation for operational control. section 1.4.2 section 12.0

Developer onboarding assistant

Engineers need answers about proprietary security protocols, deployment procedures, and compliance requirements, but the documentation is difficult to navigate. DakkaBot retrieves relevant company material and supports natural-language querying, helping newcomers find guidance faster. section 12.2

Figure 12.1 The mental map where we’re focusing on building our RAG application (12)

Use the architecture map to relate DakkaBot’s document, retrieval, prompt, safety, and evaluation components to Kubernetes, CI/CD, and experiment-tracking infrastructure. section 12.0

How Machine Learning Platform Engineering covers it

How deep it goes

Covers ingestion through embeddings and RAG, then extends into guardrails, cost optimization, and semantic evaluation.

Examples

Uses security protocols, deployment procedures, compliance requirements, guides, API documentation, runbooks, and architectural decision records.

What it assumes

Assumes established MLOps foundations while extending them with document processing, semantic search, prompt management, and LLM safety practices.

What you can do after

Build it from scratch as a deployed RAG service with ingestion, retrieval, generation, safety, cost, and evaluation components.

How it connects

Connects RAG with semantic search, vector databases, prompt engineering, safety guardrails, MLOps, and LLMOps.

liveBookmanning.com

Read Machine Learning Platform Engineering on liveBook Get the book at manning.com

This explainer was generated by Manning's knowledge graph from the text of Machine Learning Platform Engineering. The section links point to the parts of the book it draws on.