Oracle Databases 23ai: Vector Search Real Time Stage
š Vector Search Support: Enables semantic and similarity-based searches using vector embeddings within the database.
š¤ AI-Native Integration: Seamlessly works with LLMs and AI models for intelligent data retrieval.
š§ Unified Data & AI: Combines relational, JSON, and vector data in one platform using standard SQL.
ā” High Performance & Scalability: Optimized for fast, scalable vector queries on enterpriseĀ data.
Target Audience of the training
Technical professionals such as Data Engineers, Machine Learning Engineers, and Developers are involved in building AI-powered or semantic search solutions.
Outcome of the training
REQUEST TO ENROL
Training Course Structure
- Ā What is vector search? Why do we need it?
- Ā Limitations of traditional keyword search
- Ā Understanding of Use cases:
- Semantic search
- Recommendation engines
- Image/audio retrieval, etc.
- Ā Understanding of key concepts:
- Vectors
- Embeddings
- Similarity metrics
- What are embeddings?
- Word embeddings: Word2Vec, GloVe
- Sentence and document embeddings: BERT, SBERT
- Vectorizing non-text data: images, audio, structured data
- Demo: Generating embeddings using Hugging Face or sentence-transformers
- Demo: Generating embeddings using Hugging Face or sentence-transformers
- Concept of nearest neighbors
- Brute force vs approximate nearest neighbor (ANN) methods
- ANN algorithms: HNSW, IVF, PQ
- Vector databases overview: FAISS, Milvus, Weaviate, Pinecone, Vespa
- Comparison: performance, scalability, production-readiness
- Choosing a model for embeddings
- Storing vectors in a database
- Index creation and tuning
- Performing search queries
- Demo: Building a vector search app using FAISS or Weaviate
- Use case 1: Semantic search engine for documents
- Use case 2: Product recommendation system
- Use case 3: Image similarity engine
- Incorporating hybrid search: keyword + vector
- Index optimization
- Batch processing vs real-time indexing
- Latency, throughput, and cost trade-offs
- Monitoring and observability in vector search systems
- Security and data privacy in embedding models
- Multimodal vector search (text + image)
- RAG (Retrieval-Augmented Generation) in LLMs
- OpenAI Embeddings, Cohere, and commercial APIs
- Future of vector search and integration with GenAI
Training Details
Training Schedule
Timings: 08:00 AM to 10:00 AM IST
Days: Saturday to Sunday
Total Duration: 10 hours
Date: 12-Jul-24
Mode: Online
REQUEST TO ENROL
Testimonials
Good training and got to learn lot of new things on VBCS . Highly recommend!
Jithin Kumar
It was wonderful training sessions. Very precise content and good coverage of all concepts with in depth knowledge. It was a great learning experience. I would definitely recommend this training for an individual who would like to pursue career in Oracle VBCS.
Naveen Reddy
Teaching sessions of Ankur in VBCS is very helpful to excel our career in VBCS and we need to make sure we practice each day class thoroughly without fail.
Shravan Reddy
Ankur Jain has extensive knowledge of VBCS, and he conducts training in a professional manner. The agenda of the training is very clear and attempts to cover all possible common scenarios,those who are looking for new learner in VBCS can consider him, needs practice from our side is also very important thing to be considered. Thanks Ankur