Majid Fekri

Majid Fekri

Co-Founder and CTO @ Moorcheh.ai
1 point

Badges

Tastemaker
Tastemaker
Gone streaking
Gone streaking
Gone streaking 5
Gone streaking 5

Maker History

  • Moorcheh
    MoorchehSearch beyond distance. discover true relevance.
    Aug 2025
  • 🎉
    Joined Product HuntAugust 4th, 2025

Forums

Introducing Moorcheh.ai

RAG without the bottlenecks.

Faster search

Better accuracy

Edge-ready from day one

Majid Fekri

5mo ago

Moorcheh.ai – Next-Gen RAG Platform for AI Assistants & Agents

Hi Product Hunt! I m Majid.
I ve been a data scientist since 2012, starting in atmospheric sciences (predicting rain & snow ) before moving into industries like ad-tech, gaming, and telecom.

Last year I founded Edge AI Innovations to build AI assistants & agents that can run on small, resource-constrained devices. Very quickly we hit the same wall many of you might have faced Retrieval-Augmented Generation (RAG) was essential for most applications, but slow and cumbersome to implement.

So we started from scratch.

We re-imagined vector search from first principles, moved beyond cosine similarity, and invented a new distance metric. The result? Faster, more accurate retrieval and a foundation for building AI products much more easily.

Majid Fekri

5mo ago

builder of Moorcheh.ai

Hi, my name is Majid,

I ve been working as a scientific programmer and data scientist since 2012. My PhD is in Atmospheric Sciences, where my research focused on predicting rain and snow. In 2017, I transitioned into data science for other industries ad-tech, gaming, telecom, MDM, and software development.

Last year, I launched my second startup, Edge AI Innovations, to explore building AI assistants and agents for small, resource-constrained devices. Along the way, we discovered that Retrieval-Augmented Generation (RAG) is essential for most applications but often becomes a major bottleneck.

From first principles, we developed Moorcheh.ai a next-generation vector search engine built on a new distance metric (moving beyond traditional cosine similarity). The results exceeded our expectations, and we ve continued to build on that foundation.

View more