Atlas Vector Search
Atlas Vector Search Use Cases:
๐ฌ Customer Chatbot
โ Question-Answering (Q-A)
๐ Ecommerce Search
๐ฅ User Recommendations
โ๏ธ Content Generation
๐ Analysis and Summary
Most Common Atlas Vector Search Use Cases:
๐๏ธ Internal Knowledge Bases
๐ Vectorized Documentation
๐ JSON File to an Embedding Model
๐ณ LangChain or LlamaIndex
Atlas Triggers:
๐ Watch for any data changes in a single view
Atlas Vector Search:
๐ Matching documents by similarity search on indexed embeddings data
๐๏ธ Queries can use a vector's metadata (date created) to filter out older content
Atlas Search:
๐ Matching keywords, chunked customer data
๐ค Fuzzy search to correct typos
๐ฎ Autocomplete (suggested search terms)
๐ Index intersection (complex ad-hoc queries)
Infrastructure
Queryable Encryption:
๐ Securing customer data
๐ Encrypt most sensitive data uniquely identifying an individual (e.g., SSN)
Multi-document ACID Transactions:
๐ Integrity of data
Atlas Global Clusters:
๐ Define single or multi-region Zones
๐ Each zone supports write and read operations from geographically local shards
โ๏ธ Configure zones to support global low-latency secondary reads
Atlas Online Archive:
๐๏ธ Data lifecycle management
๐ฝ Automatically send outdated data from active databases into lower-cost cloud object storage
๐พ Keeping data accessible for querying
๐ฏ 9.995% uptime SLA
Distributed Architecture with Elastic Scale:
๐ Dynamically adjust database capacity
๐ Based on application demand (e.g., shopping seasonality, sales promotions)
Product Catalogs
MongoDB Product Catalogs:
๐ฆ Diversity of different products
๐ค Benefit from flexible document data model
Challenges - Keyword Search:
๐ค Without extensive and laborious synonym mapping
๐ฒ e.g., mapping bikes to cycling or sneakers to trainers
Challenges - Recommendations:
๐ง Write complex rules-based engines to get specialized and scarce data
Solution - Product Catalog with Vector Embeddings:
๐ Semantic meaning of products in the catalog
๐ค Understand similarities and relationships between products
Benefits:
๐ Search experience more intelligent & predictive
๐ Track user click-through rates
๐ฐ Sales conversions from search results
More MongoDB Features:
๐ฐ๏ธ Time Series Collections: Ingest and store high-velocity, click-streams
๐ Atlas Charts: Live visualizations of results, continuously tune and optimize business
Editor
Danny Chan, specialty of FSI and Serverless
Kenny Chan, specialty of FSI and Machine Learning
Top comments (0)