@hardiksehgal

hardik sehgal

@hardiksehgalJoined April 2025

About

Nothing here yet.

Available for

Nothing here yet.

hardik sehgal's blogs

Decoding AI jargons with Humansaiforhumans.hashnode.dev1 post

Understanding RAGunderstanding-rag.hashnode.dev1 post

Parallel Query (Fan Out) Retrievalparallel-query-fan-out-retrieval.hashnode.dev1 post

Reciprocal Rank Fusionreciprocal-rank-fusion.hashnode.dev1 post

Query Decompositionquery-decomposition.hashnode.dev1 post

Articles Comments

Recently published

HShardik sehgalworking-of-tcp.hashnode.devJan 23 · 4 min read

TCP Working: 3-Way Handshake & Reliable Communication

Before understanding the whole concept of TCP let’s first understand what were some problems that people faced while Data Transfer. Haha and yes this is way before Air Drop. TCP means Transmission Control Protocol. In the name only we got protocol wh...

HShardik sehgalhypothetical-document-embeddings-hyde.hashnode.devApr 22, 2025 · 3 min read

HyDE Query Translation - Hypothetical Document Embeddings

HyDE, or Hypothetical Document Embeddings, is a technique used in Retrieval-Augmented Generation (RAG). It involves generating a hypothetical document or answer from a question using a large language model (LLM). This is then converted into embedding...

HShardik sehgalsequential-query-decomposition.hashnode.devApr 22, 2025 · 4 min read

Sequential query decomposition with context accumulation

Sequential Query Decomposition with Context Accumulation is a method where a complex user query is broken down into smaller sub-questions. Each sub-question is answered one at a time using retrieved context. The answers are accumulated step-by-step, ...

HShardik sehgalquery-decomposition.hashnode.devApr 22, 2025 · 2 min read

What is Query Decomposition?

Instead of treating a user query as one solid question, we break it into smaller, useful pieces — either by making it: Less abstract (more specific) More abstract (higher-level/general) This helps the system retrieve better chunks from your docum...

HShardik sehgalreciprocal-rank-fusion.hashnode.devApr 22, 2025 · 3 min read

📚 What is Reciprocal Rank Fusion (RRF)?

Reciprocal Rank Fusion (RRF) is a technique used in Retrieval-Augmented Generation (RAG) systems to combine multiple ranked lists of documents into a single, improved ranking. It's particularly useful for RAG-Fusion, which enhances RAG by generating ...

hardik sehgal

About

Available for

hardik sehgal's blogs

Recently published

TCP Working: 3-Way Handshake & Reliable Communication

HyDE Query Translation - Hypothetical Document Embeddings

Sequential query decomposition with context accumulation

What is Query Decomposition?

📚 What is Reciprocal Rank Fusion (RRF)?

Search Hashnode

hardik sehgal

About

Available for

hardik sehgal's blogs

Recently published

TCP Working: 3-Way Handshake & Reliable Communication

HyDE Query Translation - Hypothetical Document Embeddings

Sequential query decomposition with context accumulation

What is Query Decomposition?

📚 What is Reciprocal Rank Fusion (RRF)?