Tag

#embedding

74 posts2 followers

Explore Hashnode

Alternatives

RONI DASOct 15 · 17 min read

I cut EmbeddingGemma 2's vectors to a sixth of their size. It got more answers right, and that was the worrying part.

A computer cannot compare meanings. It can compare numbers. So if you want a search box that understands "why can't you make good tea at the top of Everest", you first need something that turns words

systemdesignacademy.hashnode.dev

Aman JainJul 20 · 7 min read

Does a Wider Embedding Space Help More on Diverse Data?

TL;DR Increasing the embedding dimension from 32 to 64 improved validation loss in both the low- and high-diversity training conditions. Most of the improvement appeared between 32 and 48 dimensions

curious-pm.hashnode.dev

Nguyễn Hữu ĐứcJul 11 · 3 min read

Building an LLM System from Scratch in Pure Python & NumPy: Architecture, Invariants, and Clean Code, Open-Source 100% Inference Engine and Thinking engine!

Hi everyone, I wanted to share the design invariants and architecture of Draco AI, a full-stack, hardware-agnostic LLM system built completely from scratch in pure Python and NumPy. The core philosoph

draco-ai.hashnode.dev

Ashaaf KhanJul 10 · 10 min read

Behind the Illusion of Perfection: Why Your RAG Pipeline Fails in Production

When Large Language Models first took the world by storm, they felt like magic. But developers quickly ran into a brick wall: these models were stuck in the past, bounded by strict "knowledge cutoffs"

ashaaf.hashnode.dev

Tanisha SoniJul 3 · 7 min read

Your LLM Doesn't Need to Know About All 47 Tools

I built a support agent that loaded 40+ tools from MCP. The agent kept getting confused, latency crept up, and my token bill wasn't pretty. The fix turned out to be a FAISS index show the LLM only the

tanisha-soni.hashnode.dev

Raghavan T MJun 16 · 5 min read

You're Not Writing Prompts. You're Giving the AI Coordinates.

Google Photos doesn't search your photos. You type "beach" into the search bar and it finds every beach photo you've ever taken — holidays, weekends, random afternoons. No tags. No folders. You never

changeofbasis.hashnode.dev

Devathi Sai LokeshJun 4 · 6 min read

RAG is Not Just Chunking + Embedding + Retrieval — Here's What Production Actually Looks Like

Most engineers learn RAG like this: Chunk → Embed → Store → Retrieve → LLM → Done. That works for a weekend project. It fails at 10M+ documents in production. Here's what production RAG actually looks

sailokeshdevathi.hashnode.dev

eyanpenMay 4 · 11 min read

Why Does Semantic Chunking Need an Embedding API?

Fixed-length chunking requires no external services, yet semantic chunking absolutely needs an Embedding API — why? The Short Answer The core idea of semantic chunking is to split text at semantic boundaries. Determining whether "two pieces of text ...

ai-is-coming.hashnode.dev

James GainesApr 28 · 10 min read

Why Vector Search Alone Fails on Regulatory Documents

TL;DR Vector search is great at finding semantically similar content, but it struggles with broad keyword-heavy queries common in regulatory and clinical document search. By combining BM25 keyword mat

blog.gainesai.com

Aashish ChapainApr 22 · 6 min read

Embeddings in RAG: The Core of Semantic Search

Retrieval-Augmented Generation (RAG) systems rely on the core idea of converting text into vectors so we can search by meaning instead of keywords. These vectors are called embeddings, and they are th

blog.chapainaashish.com.np

Search Hashnode

Explore Hashnode

Trending tags this week

I cut EmbeddingGemma 2's vectors to a sixth of their size. It got more answers right, and that was the worrying part.

Does a Wider Embedding Space Help More on Diverse Data?

Building an LLM System from Scratch in Pure Python & NumPy: Architecture, Invariants, and Clean Code, Open-Source 100% Inference Engine and Thinking engine!

Behind the Illusion of Perfection: Why Your RAG Pipeline Fails in Production

Your LLM Doesn't Need to Know About All 47 Tools

You're Not Writing Prompts. You're Giving the AI Coordinates.

RAG is Not Just Chunking + Embedding + Retrieval — Here's What Production Actually Looks Like

Why Does Semantic Chunking Need an Embedding API?

Why Vector Search Alone Fails on Regulatory Documents

Embeddings in RAG: The Core of Semantic Search