The people dissing Deedy's main claims are clowns. Fight me. Sarvam AI did some amazing work. And yet Indian companies and devs don't care — that's an us problem, not on Sarvam.

Introducing Sarvam Samvaad 🚀 Sarvam Samvaad, our Conversational AI Platform, is designed to help enterprises build, test, and launch AI agents fluent in 11 Indian languages. ▪️Power interactions across telephone, WhatsApp, web, and apps ▪️Handle complex phrases,

What are some of the most prominent techniques to ensure/validate authenticity of LLMs ? one of them is watermarking LLMs. any other techniques do we know ?

Recently read the concepts around radix attention. Its fascinating to understand how these optimizations work under the hood. vanilla kv cache to paged attention to radix attention

Evaluation in Gen AI systems is never an aftermath thought, instead it should be a first principle. My series on Gen AI system evaluations 1. Foundational Principles In Evaluation open.substack.com/pub/deepakkuma… #GenerativeAI #Evaluation

I'm writing about evaluations in Generative AI systems. Do keep an eye on my substack posts open.substack.com/pub/deepakkuma… #generativeai #evaluation

When people get hurt, they learn to hate... When people hurt others, they become hated and racked with guilt. But knowing that pain allows people to be kind. Pain allows people to grow... and how you grow is up to you. - Jiraiya

This is so cool.

Dive into Philosophy, Reasoning, and Uncertainty in Modern AI open.substack.com/pub/deepakkuma…

I just published Different Aggregation Strategies in test-time scaling for Reasoning Models medium.com/p/different-ag…

Looks interesting

Do give it a read “Modular Prompt Design-Building Blocks Over Monoliths (Part-1)” medium.com/@deepakkumar05…

If you are interested in prompt engineering and especially how to design them for production systems, then pls check out my articles in medium medium.com/@deepakkumar05…

Do give it a read (hallucinations) open.substack.com/pub/bytenbento…

Can we collaborate here and may be leverage all the EIPs we have already in Apache Camel and see how can we use them in LangChain .

If you are interested to know about efficient llm inference methods (eg Radix Attention) then please check out this article medium.com/p/radix-attent…

I write about some internal maths and tech behind gen ai like optimizations, matrix multiplications etc. if you are also curious to learn internal details of these systems, then keep an eye on this medium publication medium.com/the-synaptic-s…

Please, do give this article a read regarding flash attention medium.com/the-synaptic-s…

Writing sth around Agentic design patterns. Now when I am linking Agentic design patterns with EIP, its like EIP in the age of Gen AI and Agents. Gregor Kudos to you and to the EIP book