{"id":3105,"date":"2023-08-08T18:50:44","date_gmt":"2023-08-08T18:50:44","guid":{"rendered":"https:\/\/salarydistribution.com\/machine-learning\/2023\/08\/08\/startup-pens-generative-ai-success-story-with-nvidia-nemo\/"},"modified":"2023-08-08T18:50:44","modified_gmt":"2023-08-08T18:50:44","slug":"startup-pens-generative-ai-success-story-with-nvidia-nemo","status":"publish","type":"post","link":"https:\/\/salarydistribution.com\/machine-learning\/2023\/08\/08\/startup-pens-generative-ai-success-story-with-nvidia-nemo\/","title":{"rendered":"Startup Pens Generative AI Success Story With NVIDIA NeMo"},"content":{"rendered":"
\n

Machine learning helped Waseem Alshikh plow through textbooks in college. Now he\u2019s putting generative AI<\/a> to work, creating content for hundreds of companies.<\/p>\n

Born and raised in Syria, Alshikh spoke no English, but he was fluent in software, a talent that served him well when he arrived at college in Lebanon.<\/p>\n

\u201cThe first day they gave me a stack of textbooks, each one a thousand pages thick, and all of it in English,\u201d he recalled.<\/p>\n

So, he wrote a program \u2014 a crude but effective statistical classifier that summarized the books \u2014 then he studied the summaries.<\/p>\n

From Concept to Company<\/b><\/h2>\n

In 2014, he shared his story with May Habib, an entrepreneur he met while working in Dubai. They agreed to create a startup that could help marketing departments \u2014 which are always pressured to do more with less \u2014 use machine learning to quickly create copy for their web pages, blogs, ads and more.<\/p>\n

\u201cInitially, the tech was not there, until transformer models<\/a> were announced \u2014 that was something we could build on,\u201d said Alshikh, the startup\u2019s CTO.<\/p>\n

<\/p>\n

\"Picture<\/p>\n

<\/a>

Writer co-founders Habib, CEO, and Alshikh, CTO.<\/figcaption><\/figure>\n

\u201cWe found a few engineers and spent almost six months building our first model, a neural network that barely worked and had about 128 million parameters,\u201d an often-used measure of an AI model\u2019s capability.<\/p>\n

Along the way, the young company won some business, changed its name to Writer<\/a> and connected with NVIDIA.<\/p>\n

A Startup Accelerated<\/b><\/h2>\n

\u201cOnce we got introduced to NVIDIA NeMo<\/a>, we were able to build industrial-strength models with three, then 20 and now 40 billion parameters, and we\u2019re still scaling,\u201d he said.<\/p>\n

NeMo is an application framework that helps companies curate their training datasets, build and customize large language models (LLMs<\/a>), and run them in production at scale. Organizations everywhere from Korea<\/a> to Sweden<\/a> are using it to customize LLMs for their local languages and industries.<\/p>\n

\u201cBefore NeMo, it took us four and a half months to build a new billion-parameter model. Now we can do it in 16 days \u2014 this is mind blowing,\u201d Alshikh said.<\/p>\n

Models Make Opportunities<\/b><\/h2>\n

In the first six months of this year, the startup\u2019s team of fewer than 20 AI engineers used NeMo to develop 10 models, each with 30 billion parameters or more.<\/p>\n

That translates into big opportunities. Hundreds of businesses now use Writer\u2019s models that NeMo customized for finance, healthcare, retail and other vertical markets.<\/p>\n

<\/p>\n

\"Writer's<\/p>\n

<\/a>

Writer\u2019s Recap tool creates written summaries from audio recordings of an interview or event.<\/figcaption><\/figure>\n

The startup\u2019s customer list includes household names like Deloitte, L\u2019Oreal, Intuit, Uber and many Fortune 500 companies.<\/p>\n

Writer\u2019s success with NeMo is just the start of the story. Dozens of other companies have already downloaded NeMo.<\/p>\n

The software will be available soon for anyone to use. It\u2019s part of NVIDIA AI Enterprise<\/a>, full-stack software optimized to accelerate generative AI workloads and backed by enterprise-grade support, security and application programming interface stability.<\/p>\n

<\/p>\n

\"Writer's<\/p>\n

<\/a>

Writer offers a full-stack platform for enterprise users.<\/figcaption><\/figure>\n

A Trillion API Calls a Month<\/b><\/h2>\n

Some customers run Writer\u2019s models on their own systems or cloud services. Others ask Writer to host the models, or they use Writer\u2019s API.<\/p>\n

\u201cOur cloud infrastructure, managed basically by two people, hosts a trillion API calls a month \u2014 we\u2019re generating 90,000 words a second,\u201d Alshikh said. \u201cWe\u2019re delivering high-quality models that compete with products from companies with larger teams and bigger budgets.\u201d<\/p>\n

<\/p>\n

\"Chart<\/p>\n

<\/a>

NVIDIA NeMo supports an end-to-end flow for generative AI from data curation to inference.<\/figcaption><\/figure>\n

Writer uses the Triton Inference Server<\/a> that\u2019s packaged with NeMo to run models in production for its customers. Alshikh reports that Triton, used by many companies running LLMs<\/a>, enables lower latency and greater throughput than alternative programs.<\/p>\n

\u201cThis means you can run a service for $20,000, instead of $100,000, so we can invest more in building meaningful features,\u201d he said.<\/p>\n

A Wide Horizon<\/b><\/h2>\n

Writer is also a member of NVIDIA Inception<\/a>, a program that nurtures cutting-edge startups. \u201cThanks to Inception, we got early access to NeMo and some amazing people who guided us through the process of finding and using the tools we need,\u201d he said.<\/p>\n

Now that Writer\u2019s text products are getting traction, Alshikh, who splits his time between homes in Florida and California, is searching the horizon for what\u2019s next. In today\u2019s broad frontier of generative AI, he sees opportunities in images, audio, video, 3D \u2014 maybe all of the above.<\/p>\n

\u201cWe see multimodality as the future,\u201d he said.<\/p>\n

Check out this page<\/a> to get started with NeMo. And learn about the early access program for multimodal NeMo here<\/a>.<\/p>\n

And if you enjoyed this story, let folks on social networks know using the following, a summary suggested by Writer:<\/p>\n

\u201cLearn how startup Writer uses NVIDIA NeMo software to generate content for hundreds of companies and rack up impressive revenues with a small staff and budget.\u201d<\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

https:\/\/blogs.nvidia.com\/blog\/2023\/08\/08\/writer-nemo-generative-ai\/<\/p>\n","protected":false},"author":0,"featured_media":3106,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[3],"tags":[],"_links":{"self":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts\/3105"}],"collection":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/comments?post=3105"}],"version-history":[{"count":0,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/posts\/3105\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/media\/3106"}],"wp:attachment":[{"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/media?parent=3105"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/categories?post=3105"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/salarydistribution.com\/machine-learning\/wp-json\/wp\/v2\/tags?post=3105"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}