Ganesh Bhat

#NVIDIA #Merlin meets the #MLOps ecosystem: building a production-ready #RecSys pipeline on cloud medium.com/nvidia-merlin/…

Google PM open-sources Always On Memory Agent, ditching vector databases for LLM-driven persistent memory. No #vectordatabase. No embeddings. Just an #LLM that reads, thinks, and writes structured memory #opensource #persistent #memory venturebeat.com/orchestration/… via @VentureBeat

New #KV #cache compaction technique cuts #LLM #memory 50x without accuracy loss venturebeat.com/orchestration/… via @VentureBeat

#Nvidia reportedly building its own #AI #agent to compete with #OpenClaw - ‘NemoClaw’ will supposedly be #opensource & designed for #enterprise use. It could potentially give AI tools more independence & do more things without direct human supervision tomshardware.com/tech-industry/…

Forcing #LLMs to be evil during #training can make them nicer in the long run technologyreview.com/2025/08/01/112…

#Databricks' #serverless #database slashes app development from months to days as companies prep for #agentic #AI venturebeat.com/data/databrick… via @VentureBeat

Made in India for Indian students ! pschool.in

An #opensource #CUDA C++ #compiler written from scratch in #C99 that takes .cu files and compiles them to #AMD #GPU #machine code, with more architectures planned. No LLVM, no dependencies, and no permission asked github.com/Zaneham/BarraC…

#Causal #analysis introduces a “what if” component. It goes beyond observation to ask how the system would react if we actively changed a variable #machinelearning towardsdatascience.com/causal-ml-for-…

#MIT's new #fine-#tuning method called self-distillation fine-tuning (#SDFT), allows models to learn directly from demonstrations & their own experiments by leveraging the inherent in-context learning abilities of modern #LLMs venturebeat.com/orchestration/… via @VentureBeat

#Google #Chrome ships #WebMCP in early preview, turning every #website into a structured #tool for #AI #agents venturebeat.com/infrastructure… via @VentureBeat

Hey @digilocker_ind, how to retrieve documents which have name prior to marriage (maiden name) as current name in Aadhar is post marriage which is different. This would be the case for most women in India.

#FalkorDB is the first queryable Property #Graph #database to leverage #sparse #matrices for representing the #adjacencymatrix in #graphs and #linearalgebra for #querying.Their goal provide the best #Knowledge #Graph for #LLM (#GraphRAG) github.com/FalkorDB/Falko…

Self-Evolving #Agents - A fully autonomous framework that evolves high-performing #language agents through #multistep co-evolution and seamless #tool integration. github.com/aiming-lab/Age…

#OpenNotebook - an #opensource #NotebookLM alternative that's powerful, #private - and #free zdnet.com/article/i-foun…

#Agentset is the #opensource platform to build, evaluate, and ship #production-ready #RAG and #agentic applications. It provides end-to-end tooling: ingestion, #vector indexing, evaluation/benchmarks, #chat playground, hosting, & a developer-friendly #api3 github.com/agentset-ai/ag…

#Claude is trained to be politically even-handed in its responses. They want to treat opposing political viewpoints with equal #depth, #engagement, & #quality of #analysis, without bias towards or against any particular ideological position anthropic.com/news/political…

🚀 Big AI updates in @GoogleColab! We're making generative AI more accessible and easier to use with two new features: 1️⃣ The google.colab.ai Python library is now available to ALL users! Free tier users now get access to Gemini flash models at no cost. 2️⃣ All users get access to a new AI Prompt Cell for easy, low-code AI interaction. Find it in the dropdown next to the "+ Code" button. See blog below 👇

Practical #LLM #Security Advice from the NVIDIA #AI #Red Team developer.nvidia.com/blog/practical…

#MinMax vs #Standard vs #Robust #Scaler: Which One Wins for #Skewed #Data? #machinelearning machinelearningmastery.com/minmax-vs-stan…