In this AWS re:Invent 2023 video, Aravind Srinivas, cofounder and CEO of Perplexity, shares how the company is reimagining search by providing personalized answers using Amazon Web Services (AWS) instead of link lists and generic results.
Perplexity, a conversational artificial intelligence (AI) company, is using AWS to create an interactive search companion that uses just a few follow-up questions to clarify your query and find the answer you’re looking for—no prompt engineering required. Perplexity uses Anthropic’s Claude 2 on Amazon Bedrock to quickly test and deploy improvements to its general question answering capabilities, provide more natural sounding answers, and build new features, like analyzing content from large files uploaded by users.
Perplexity worked with the AWS Startups team to simplify debugging large model training on Scuderia and increase the training throughput of cutting-edge open-source large language models (LLMs) using Amazon SageMaker, Amazon EC2 P4 instances, and AWS Elastic Fabric Adapter. With this support, Perplexity improved inferencing and debugging of large model training while handling distributed capacity efficiently. This improved the helpfulness and accuracy of its in-house models, and enabled Perplexity to create an LLM with no knowledge cutoff that’s fully hosted on AWS. Perplexity is innovating to become earth’s most knowledge centric company using the power of generative AI and AWS.
Reimagining the future of search by trying to take us from ten blue links, to personalized answers that cut through the noise and gets to exactly what you want… Our product is complex to run, and a hard problem to solve. That’s why we decided to go all in on AWS, testing frontier models including Anthropic’s Claude 2 on Amazon Bedrock.”
Aravind Srinivas
Cofounder and CEO, Perplexity
About Perplexity AI
Perplexity is a search engine platform that uses artificial intelligence to provide large language models and search engines. The company's platform enables the development of beneficial artificial general intelligence, as well as an open-source environment that is accessible to the public, allowing clients to develop skills and knowledge inartificial intelligence.
Customer Speaker: Aravind Srinivas
Aravind Srinivas is an accomplished entrepreneur and visionary leader, currently serving as the CEO of Perplexity AI. With a deep passion for technology and a relentless pursuit of innovation, Aravind has made significant contributions to the field of artificial intelligence and machine learning. Aravind's journey in the world of technology began early in his life, driven by an insatiable curiosity and a natural aptitude for problem-solving. He received his PhD in Computer Science from Berkeley, where he honed his skills and gained a solid foundation in cutting-edge technologies. After receiving his PhD in Computer Science from UC Berkeley, Aravind embarked on a career focusing in artificial intelligence. Aravind's contributions to the tech industry have earned him numerous accolades and recognition. He has been featured in leading publications, highlighting his achievements and the transformative impact of Perplexity AI's solutions.
More Software & Internet Customer Stories
Total results: 808
no items found
-
United States
Salesforce Migrates Its Marketing Cloud to AWS
Learn how CRM provider Salesforce is modernizing its Marketing Cloud solution to save money and deploy faster on AWS. -
United States
Salesforce Makes Its Observability Platform More Scalable
Learn how Salesforce in the software industry modernized its on-premises infrastructure by migrating to AWS. -
Other
Salesforce hosts its 2023 NYC Summit on AWS
This case study provides an overview of the 2023 Salesforce World Tour event held in New York City, which showcased how AWS and Salesforce are collaborating to drive innovation. -
Americas
Millions Saved without Compromise Using Amazon S3 with Salesforce
Using services like Amazon S3 Intelligent-Tiering, Amazon EMR, Amazon EC2, and AWS MAP, Salesforce migrated to AWS, and saves millions of dollars annually while increasing the performance, efficiency, and scalability of its data lake.
More to Explore
# Screens < wide
# screens = wide
