My X feed has been full of Jev as an embedder, so I got curious. Can good decisions make good coordinates for search? Turns out, they don't. That said, Jev is an outstanding reranker and post-filter.
ramsriharsha.substack.com/p/…
The world's best coding agent rejected vector search and chose grep. A tool from 1973.
Your agent's memory doesn't learn. It never has.
I wrote about why that matters and what comes next.
open.substack.com/pub/ramsri…
Ram Sriharsha retweeted
🚀 Announcing llama-text-embed-v2 in Pinecone Inference
An embedding model built on @meta Llama 3.2 1B matches or exceeds OpenAI's Large embedding retrieval quality with 10x lower p99 latency
ALT bar chart comparing NCDG@10 across MIRACL, FiQA, NQ, HotpotQA
Ram Sriharsha retweeted
Easily generate one billion embeddings in under a day... and with limited budget! How? With Ray Data from @anyscalecompute and Pinecone! Don't miss this session during the #DataAISummit tomorrow. Catch it in-person or online. databricks.com/dataaisummit/…
Ram Sriharsha retweeted
I am thrilled to share our latest "Generation AI" podcast with @halfabrane, who I've known since back in the days at Hortonworks. Ram is VPE at @pinecone and shares his insights about vector databases and the what the future holds for Generative AI.
open.spotify.com/episode/1eB…
Ram Sriharsha retweeted
We are deeply honoured to partner with Edo (@EdoLiberty) , Bob Wiederhold, Greg (@grigoriy_kogan) , Ram (@halfabrane), and the rest of the team on their vision of becoming the memory layer for next gen AI applications.
Ram Sriharsha retweeted
.@TimT toasts @pinecone on their milestone $100m Series B & $750m valuation (led by @a16z), describes how Pinecone became core to the modern AI data stack, and roasts founders @EdoLiberty & @Halfabrane in this not-very-VC blog post (+ Avengers analogies!) mnlo.vc/TimEdoRam
i really enjoyed this conversation w./ @QuinnyPig . Thanks so much for hosting me and for the great questions!
Weekend listening is officially assigned. Listen to our VP of Engineering and R&D, @halfabrane, as he joins @QuinnyPig of Screaming in the Cloud to discuss why our customer adoption has seen such a rapid rise, and much more. hubs.ly/Q01tVNBf0
Ram Sriharsha retweeted
This week on #TheDataExchangePod, @halfabrane, VP of Engineering and R&D @pinecone. We discuss vector databases in general, and Pinecone's new proprietary storage engine (written in the Rust programming language #rustlang)
thedataexchange.media/a-new-…
Ram Sriharsha retweeted
🚨Big news: We just released three huge features...
🚀 p2 pods - up to 10x faster search and higher throughput
📈 Vertical scaling - with zero downtime
🗄️ Collections - save and reuse your vector data
Read the full announcement to learn more:
pinecone.io/learn/faster-eas…
Ram Sriharsha retweeted
OnlineSTL: Scaling Time Series Decomposition by 100x
deepai.org/publication/onlin…
by @abhinavm24 et al.
#AnomalyDetection #ComputerScience
Ram Sriharsha retweeted
Vol:15 No:7 → OnlineSTL: Scaling Time Series Decomposition by 100x vldb.org/pvldb/vol15/p1417-m…
Ram Sriharsha retweeted
I interviewed Ram Sriharsha of pinecone.io for this new episode of @seradio. He explained some interesting machine learning basics before diving into their vector search database offering. @pinecone @halfabrane se-radio.net/2022/01/episode…
super excited to announce smle-beta, a reimagining of the data science experience on top of splunk to help data scientists put together machine learning workflows and operationalize them with ease on Splunk
We're pretty sure we've got a little something that will bring a SM(I)LE to your face. Introducing the beta of Splunk Machine Learning Environment, a world where SPL and data scientists come together on the same platform, for optimal speed and insights together. #splunkconf20
Ram Sriharsha retweeted
Learn how @splunk uses @ApacheFlink to build machine learning systems and how to support #ML on streams with @halfabrane at #flinkforward San Francisco. bit.ly/3ca7VFk #analytics #streamprocessing
Ram Sriharsha retweeted
More details on the Databricks Unified Analytics Platform for Genomics, which we launched today to simplify and accelerate genomic data processing: databricks.com/blog/2018/06/…
Ram Sriharsha retweeted
.@databricks @halfabrane & Frank share what you can do today w/ #UnifiedAnalyticsPlatform for #Genomics @SparkAISummit #ApacheSpark – at San Francisco, CA
very excited to share the work we have been doing to scale genomics on the cloud. databricks.com/blog/2018/06/…
Ram Sriharsha retweeted
Spark Geospatial Analytics with Ram Sriharsha @halfabrane @ApacheSpark @Databricks #Data #Magellan softwareengineeringdaily.com…
Ram Sriharsha retweeted
Going to #StrataConf? Come hear Databricks @halfabrane talk: 'Magellan: Scalable and fast geospatial analytics' dbricks.co/2oPwgt9