Skip to content

Commit 448f518

Browse files
authored
Merge pull request #56 from 567-labs/improve-index-navigation
Improve docs/index.md as comprehensive table of contents
2 parents ac3fb16 + 2231441 commit 448f518

File tree

1 file changed

+129
-87
lines changed

1 file changed

+129
-87
lines changed

docs/index.md

Lines changed: 129 additions & 87 deletions
Original file line numberDiff line numberDiff line change
@@ -10,15 +10,15 @@ date: 2025-04-10
1010

1111
## Data-Driven Product Development for AI Applications
1212

13-
_A systematic approach to building self-improving AI systems_
13+
*A systematic approach to building self-improving AI systems*
1414

1515
!!! abstract "About This Book"
16-
This book provides a structured approach to evolving Retrieval-Augmented Generation (RAG) from a technical implementation into a continuously improving product. You'll learn to combine product thinking with data science principles to create AI systems that deliver increasing value over time.
16+
This book provides a structured approach to evolving Retrieval-Augmented Generation (RAG) from a technical implementation into a continuously improving product. You'll learn to combine product thinking with data science principles to create AI systems that deliver increasing value over time.
17+
18+
Most teams focus on the latest models and algorithms while missing the fundamentals: understanding their data, measuring performance, and systematically improving based on user feedback.
1719

1820
## The RAG Improvement Flywheel
1921

20-
At the core of this book is the RAG improvement flywheel - a continuous cycle that transforms user interactions into product enhancements.
21-
2222
```mermaid
2323
graph TD
2424
A[Synthetic Data & Evaluation] --> B[Learning from Evaluations]
@@ -33,129 +33,171 @@ graph TD
3333
style E fill:#dfd,stroke:#333,stroke-width:2px
3434
```
3535

36-
!!! tip "Beyond Technical Implementation"
37-
This book goes beyond teaching you how to implement RAG. It shows you how to think about RAG as a product that continuously evolves to meet user needs and deliver business value.
38-
39-
## Chapters
36+
## Workshop Series
4037

4138
### [Introduction: Beyond Implementation to Improvement](workshops/chapter0.md)
39+
Shifting from technical implementation to product-focused continuous improvement. Understanding RAG as a recommendation engine and the improvement flywheel.
4240

43-
Understand why systematic improvement matters and how to approach RAG as a product rather than just a technical implementation.
41+
### [Chapter 1: Kickstarting the Data Flywheel with Synthetic Data](workshops/chapter1.md)
42+
Common pitfalls in AI development, leading vs. lagging metrics, precision and recall for retrieval evaluation, and synthetic data generation techniques.
4443

45-
### [Chapter 1: Starting the Flywheel](workshops/chapter1.md)
44+
### [Chapter 2: Converting Evaluations into Training Data for Fine-Tuning](workshops/chapter2.md)
45+
Why generic embeddings fall short, converting evaluation examples into few-shot prompts, contrastive learning, and re-rankers as cost-effective strategies.
4646

47-
Learn how to overcome the cold-start problem, establish meaningful metrics, and create a data foundation that drives product decisions.
47+
### Chapter 3: User Experience and Feedback Collection
4848

49-
### [Chapter 2: From Evaluation to Enhancement](workshops/chapter2.md)
49+
#### [Chapter 3.1: Feedback Collection](workshops/chapter3-1.md)
50+
Making feedback visible and engaging (increasing rates from <1% to >30%), proven copy patterns, and enterprise feedback collection through Slack integrations.
5051

51-
Transform evaluation insights into concrete product improvements through fine-tuning, re-ranking, and targeted enhancements.
52+
#### [Chapter 3.2: Streaming and Interstitials](workshops/chapter3-2.md)
53+
Psychology of waiting, implementing streaming responses for 30-40% higher feedback collection, and meaningful interstitials.
5254

53-
### [Chapter 3: The User Experience of AI](workshops/chapter3-1.md)
55+
#### [Chapter 3.3: Quality Improvements](workshops/chapter3-3.md)
56+
Interactive citations, chain of thought reasoning for 8-15% accuracy improvements, and validation patterns reducing errors by 80%.
5457

55-
Design interfaces that both delight users and gather valuable feedback, creating a virtuous cycle of improvement.
58+
### Chapter 4: Understanding Your Users
5659

57-
### [Chapter 4: Understanding Your Users](workshops/chapter4-1.md)
60+
#### [Chapter 4.1: Topic Modeling and Analysis](workshops/chapter4-1.md)
61+
Moving from individual feedback to systematic pattern identification and transforming "make the AI better" into specific priorities.
5862

59-
Segment users and queries to identify high-value opportunities and create targeted improvement strategies.
63+
#### [Chapter 4.2: Prioritization and Roadmapping](workshops/chapter4-2.md)
64+
Impact/effort prioritization using 2x2 frameworks and building strategic roadmaps based on user behavior patterns.
6065

61-
### [Chapter 5: Building Specialized Capabilities](workshops/chapter5-1.md)
66+
### Chapter 5: Building Specialized Retrieval
6267

63-
Develop purpose-built solutions for different user needs spanning documents, images, tables, and structured data.
68+
#### [Chapter 5.1: Understanding Specialized Retrieval](workshops/chapter5-1.md)
69+
Why monolithic approaches reach limits, two complementary strategies (extracting metadata vs. creating synthetic text), and two-level measurement.
6470

65-
### [Chapter 6: Unified Product Architecture](workshops/chapter6-1.md)
71+
#### [Chapter 5.2: Implementing Multimodal Search](workshops/chapter5-2.md)
72+
Advanced document retrieval, image search challenges, table search approaches, SQL generation, and RAPTOR hierarchical summarization.
6673

67-
Create a cohesive product experience that intelligently routes to specialized components while maintaining a seamless user experience.
74+
### Chapter 6: Unified Architecture
6875

69-
### [Key Takeaways: Product Principles for AI Applications](misc/what-i-want-you-to-takeaway.md)
76+
#### [Chapter 6.1: Query Routing Foundations](workshops/chapter6-1.md)
77+
The API mindset, organizational structure, evolution from monolithic to modular architecture, and performance formula.
7078

71-
Core principles that will guide your approach to building AI products regardless of how the technology evolves.
79+
#### [Chapter 6.2: Tool Interfaces and Implementation](workshops/chapter6-2.md)
80+
Designing tool interfaces, router implementation using structured outputs, and dynamic example selection.
7281

73-
## Talks and Presentations
82+
#### [Chapter 6.3: Performance Measurement](workshops/chapter6-3.md)
83+
Measuring tool selection effectiveness, dual-mode UI, user feedback as training data, and creating improvement flywheel.
7484

75-
Explore insights from industry experts and practitioners through our collection of talks, lightning lessons, and presentations:
85+
## Expert Talks
7686

77-
### [Featured Talks](talks/index.md)
87+
### Foundation and Evaluation
7888

79-
- **[Fine-tuning Re-rankers and Embedding Models for Better RAG Performance](talks/fine-tuning-rerankers-embeddings-ayush-lancedb.md)** - Practical approaches to enhancing retrieval quality (Ayush from LanceDB)
80-
- **[RAG Anti-patterns in the Wild](talks/rag-antipatterns-skylar-payne.md)** - Common mistakes and how to fix them (Skylar Payne)
81-
- **[Semantic Search Over the Web with Exa](talks/semantic-search-exa-will-bryk.md)** - Building AI-first search engines (Will Bryk)
82-
- **[Understanding Embedding Performance through Generative Evals](talks/embedding-performance-generative-evals-kelly-hong.md)** - Custom evaluation methodologies (Kelly Hong)
83-
- **[Online Evals and Production Monitoring](talks/online-evals-production-monitoring-ben-sidhant.md)** - Monitoring AI systems at scale (Ben Hylak & Sidhant Bendre)
89+
**[Building Feedback Systems](talks/zapier-vitor-evals.md)** - Vitor (Zapier)
90+
Simple UX changes increased feedback collection 4x. Key insight: specific questions like "Did this run do what you expected?" dramatically outperform generic prompts.
8491

85-
[View all talks →](talks/index.md)
92+
**[Text Chunking Strategies](talks/chromadb-anton-chunking.md)** - Anton (ChromaDB)
93+
Why chunking remains critical even with infinite context windows. Default chunking strategies in popular libraries often produce terrible results.
8694

87-
## For Product Leaders, Engineers, and Data Scientists
95+
**[Embedding Performance Evaluation](talks/embedding-performance-generative-evals-kelly-hong.md)** - Kelly Hong
96+
Model rankings on custom benchmarks often contradict MTEB rankings - public benchmark performance doesn't guarantee real-world success.
8897

89-
!!! info "What You'll Learn"
90-
**For Product Leaders**
91-
- How to establish metrics that align with business outcomes
92-
- Frameworks for prioritizing AI product improvements
93-
- Approaches to building product roadmaps for RAG applications
94-
- Methods for communicating AI improvements to stakeholders
95-
**For Engineers**
96-
- Implementation patterns that facilitate rapid iteration
97-
- Architectural decisions that enable continuous improvement
98-
- Techniques for building modular, specialized capabilities
99-
- Approaches to technical debt management in AI systems
100-
101-
**For Data Scientists**
102-
- Methods for creating synthetic evaluation datasets
103-
- Techniques for segmenting and analyzing user queries
104-
- Frameworks for measuring retrieval effectiveness
105-
- Approaches to continuous learning from user interactions
98+
### Training and Fine-Tuning
10699

107-
## Quick Wins: High-Impact RAG Improvements
100+
**[Enterprise Search Fine-tuning](talks/glean-manav.md)** - Manav (Glean)
101+
Custom embedding models achieve 20% improvements through continuous learning. Smaller, fine-tuned models often outperform larger general-purpose models.
108102

109-
Based on real-world implementations, here are proven improvements you can implement quickly:
103+
**[Fine-tuning Re-rankers](talks/fine-tuning-rerankers-embeddings-ayush-lancedb.md)** - Ayush (LanceDB)
104+
Re-rankers provide 12-20% retrieval improvement with minimal latency penalty - "low-hanging fruit" for RAG optimization.
110105

111-
!!! success "Top 5 Quick Wins"
112-
1. **Change Feedback Copy**
113-
- Replace "How did we do?" with "Did we answer your question?"
114-
- **Impact**: 5x increase in feedback collection
115-
- **Effort**: 1 hour
116-
117-
2. **Use Markdown Tables**
118-
- Format structured data as markdown tables instead of JSON/CSV
119-
- If tables are complex, represent it in XML
120-
- **Impact**: 12% better lookup accuracy
121-
- **Effort**: 2-4 hours
122-
123-
3. **Add Streaming Progress**
124-
- Show "Searching... Analyzing... Generating..." with progress
125-
- Stream the response as it's being generated when possible
126-
- **Impact**: 45% reduction in perceived latency
127-
- **Effort**: 1 sprint
128-
129-
4. **Implement Page-Level Chunking**
130-
- For documentation, respect page boundaries, and use page-level chunking. Humans tend to create semantically coherent chunks at the page level.
131-
- **Impact**: 20-30% better retrieval for docs
132-
- **Effort**: 1 day
106+
### Production and Monitoring
107+
108+
**[Production Monitoring](talks/online-evals-production-monitoring-ben-sidhant.md)** - Ben & Sidhant
109+
Traditional error monitoring doesn't work for AI since there's no exception when models produce bad outputs.
110+
111+
**[RAG Anti-patterns](talks/rag-antipatterns-skylar-payne.md)** - Skylar Payne
112+
90% of teams adding complexity see worse performance. Silent failures in document processing can eliminate 20%+ of corpus without detection.
133113

134-
!!! tip "Medium-Term Improvements (2-4 weeks)"
114+
### Specialized Retrieval
115+
116+
**[Agentic RAG](talks/colin-rag-agents.md)** - Colin Flaherty
117+
Simple tools like grep and find outperformed sophisticated embedding models due to agent persistence and course-correction capabilities.
118+
119+
**[Better Data Processing](talks/reducto-docs-adit.md)** - Adit (Reducto)
120+
Hybrid computer vision + VLM pipelines outperform pure approaches. Even 1-2 degree document skews dramatically impact extraction quality.
121+
122+
**[Multi-Modal Retrieval](talks/superlinked-encoder-stacking.md)** - Daniel (Superlinked)
123+
LLMs fundamentally can't understand numerical relationships. Use mixture of specialized encoders for different data types.
124+
125+
**[Lexical Search](talks/john-lexical-search.md)** - John Berryman
126+
Semantic search struggles with exact matching and specialized terminology. Lexical search provides efficient filtering and rich metadata.
127+
128+
## Quick Wins: High-Impact RAG Improvements
129+
130+
!!! success "Top Quick Wins"
131+
1. **Change Feedback Copy**: Replace "How did we do?" with "Did we answer your question?"
132+
2. **Use Markdown Tables**: Format structured data as markdown tables instead of JSON/CSV
133+
3. **Add Streaming Progress**: Show "Searching... Analyzing... Generating..." with progress
134+
4. **Page-Level Chunking**: For documentation, respect page boundaries for better retrieval
135+
136+
!!! tip "Medium-Term Improvements"
135137
- **Fine-tune embeddings**: $1.50 and 40 minutes for 6-10% improvement
136138
- **Add re-ranker**: 15-20% retrieval improvement
137139
- **Build specialized tools**: 10x better for specific use cases
138-
- **Implement contextual retrieval**: 30% better context understanding
139-
- **Create Slack feedback integration**: 5x more enterprise feedback
140+
- **Slack feedback integration**: 5x more enterprise feedback
141+
142+
## How to Use This Resource
143+
144+
**For Beginners**: Start with the [Introduction](workshops/chapter0.md), then work through chapters sequentially.
145+
146+
**For Quick Wins**: Jump to the [Quick Wins section](#quick-wins-high-impact-rag-improvements) for immediate improvements.
140147

141-
!!! info "Learn from the Experts"
142-
Before implementing, learn from these practical talks:
143-
- [**RAG Anti-patterns in the Wild**](talks/rag-antipatterns-skylar-payne.md) - Common mistakes across industries and how to fix them
144-
- [**Document Ingestion Best Practices**](talks/reducto-docs-adit.md) - Production-ready parsing for tables, PDFs, and complex documents
148+
**For Specific Problems**: Check the [FAQ](office-hours/faq.md) for answers to common questions.
149+
150+
**For Complete Implementation**: Follow the full workshop series from Chapter 1 through 6.3.
151+
152+
## Key Insights
153+
154+
**Most Important**: Teams that iterate fastest on data examination consistently outperform those focused on algorithmic sophistication.
155+
156+
**Most Underutilized**: Fine-tuning embeddings and re-rankers are more accessible and impactful than most teams realize.
157+
158+
**Biggest Mistake**: 90% of teams add complexity that makes their RAG systems worse. Start simple, measure everything, improve systematically.
159+
160+
## Frequently Asked Questions
161+
162+
- **"Is knowledge graph RAG production ready?"** Probably not. [See why →](office-hours/faq.md#is-knowledge-graph-rag-production-ready-by-now-do-you-recommend-it)
163+
- **"How do we handle time-based queries?"** Use PostgreSQL with pgvector-scale. [Learn more →](office-hours/faq.md#how-do-we-introduce-a-concept-of-time-and-vector-search-to-answer-questions-like-whats-the-latest-news-without-needing-to-move-to-a-graph-database)
164+
- **"Should we use DSPy for prompt optimization?"** It depends. [See when →](office-hours/faq.md#what-is-your-take-on-dspy-should-we-use-it)
165+
- **"Would you recommend ColBERT models?"** Test against your baseline first. [See approach →](office-hours/faq.md#would-you-recommend-using-colbert-models-or-other-specialized-retrieval-approaches)
166+
167+
[Browse All FAQ](office-hours/faq.md){ .md-button } [View Office Hours](office-hours/index.md){ .md-button }
168+
169+
## For Product Leaders, Engineers, and Data Scientists
170+
171+
!!! info "What You'll Learn"
172+
**For Product Leaders**: Establish metrics that align with business outcomes, prioritization frameworks, and roadmapping approaches
173+
174+
**For Engineers**: Implementation patterns for rapid iteration, architectural decisions, and modular capabilities
175+
176+
**For Data Scientists**: Synthetic evaluation datasets, query segmentation techniques, and continuous learning approaches
177+
178+
## Navigate by Topic
179+
180+
**Evaluation & Metrics**: [Chapter 1](workshops/chapter1.md)[Kelly Hong Talk](talks/embedding-performance-generative-evals-kelly-hong.md)[Vitor Zapier Talk](talks/zapier-vitor-evals.md)
181+
182+
**Fine-tuning & Training**: [Chapter 2](workshops/chapter2.md)[Ayush LanceDB Talk](talks/fine-tuning-rerankers-embeddings-ayush-lancedb.md)[Manav Glean Talk](talks/glean-manav.md)
183+
184+
**User Experience**: [Chapter 3 Series](workshops/chapter3-1.md)[Streaming Guide](workshops/chapter3-2.md)[Quality Improvements](workshops/chapter3-3.md)
185+
186+
**Architecture & Routing**: [Chapter 6 Series](workshops/chapter6-1.md)[Query Routing](talks/query-routing-anton.md)[Multi-modal Retrieval](talks/superlinked-encoder-stacking.md)
145187

146188
## About the Author
147189

148-
Jason Liu brings practical experience from his work at Facebook, Stitch Fix, and as a consultant for companies like HubSpot, Zapier, and many others. His background spans computer vision, recommendation systems, and RAG applications across diverse domains.
190+
Jason Liu brings practical experience from Facebook, Stitch Fix, and as a consultant for companies like HubSpot, Zapier, and many others. His background spans computer vision, recommendation systems, and RAG applications across diverse domains.
149191

150192
!!! quote "Author's Philosophy"
151-
"The most successful AI products aren't the ones with the most sophisticated models, but those built on disciplined processes for understanding users, measuring performance, and systematically improving. This book will show you how to create not just a RAG application, but a product that becomes more valuable with every interaction."
193+
"The most successful AI products aren't the ones with the most sophisticated models, but those built on disciplined processes for understanding users, measuring performance, and systematically improving."
194+
195+
---
152196

153197
## Getting Started
154198

155199
Begin your journey by reading the [Introduction](workshops/chapter0.md) or jump directly to [Chapter 1](workshops/chapter1.md) to start building your evaluation framework and data foundation.
156200

157-
---
158-
159-
IF you want to get discounts and 6 day email source on the topic make sure to subscribe to
201+
If you want to get discounts and 6 day email source on the topic make sure to subscribe to
160202

161-
<script async data-uid="010fd9b52b" src="https://fivesixseven.kit.com/010fd9b52b/index.js"></script>
203+
<script async data-uid="010fd9b52b" src="https://fivesixseven.kit.com/010fd9b52b/index.js"></script>

0 commit comments

Comments
 (0)