<\/strong><\/h3>\n
Dessa has distilled its experience engineering ML for the world\u2019s largest companies into powerful ML products focused on development and production: <\/strong><\/em>Try Atlas for free<\/strong><\/em><\/a><\/h4>\nToronto, ON (Oct. 21, 2019) \u2013 Dessa, a company with world-renowned excellence in machine learning, has announced the launch of its products Atlas and Orbit, which together make up Foundations, an end-to-end platform for the development and production of ML at scale.<\/p>\n
The vision for both products is simple: with powerful tools developing real-world ML and maintaining it after production, we want Atlas and Orbit to give engineers around the world the ability to build impactful applications with ML faster than ever before.<\/p>\n
The launch of Atlas and Orbit marks the company\u2019s biggest product launch to date, coming after Dessa engineers shared their AI recreation of Joe Rogan\u2019s voice, developed using Atlas.<\/p>\n
Product Details:<\/h4>\n\n- Atlas:<\/strong> Tools for rapidly developing ML at scale. Enables ML engineers to manage 1000s of experiments concurrently, save 8x on compute costs and deploy models with a single click<\/li>\n
- Orbit:<\/strong> Tools for managing ML models in production. Enables ML engineers to overcome costly production challenges, with features for model performance monitoring, maintenance, automatic training and A\/B testing<\/li>\n<\/ul>\n
\u201cAtlas transforms the time-consuming and frustrating process of managing experiments into something that\u2019s easy and even enjoyable, also allowing for insights you might not have otherwise,\u201d said Alex Krizhevsky, Principal ML Architect at Dessa.<\/p>\n
The design of both products was informed by the company\u2019s experience deploying ML for the world\u2019s largest organizations, including Fortune 500 companies. Stephen Piron, the company\u2019s Co-CEO remarks that \u201cbefore Foundations, helping these companies build and deploy their first AI systems was never easy. It often felt like we had to drag these systems kicking and screaming into production. But everything we\u2019ve learned has been baked into Atlas and Orbit. We hope offering these tools to the world will be truly game-changing.\u201d<\/p>\n
Teams wanting to test the products themselves can register for a free trial of Atlas<\/a>, also featuring Google Cloud credits and a text-generation tutorial. Demos for Atlas and Orbit are also available.<\/p>\nAbout Dessa<\/b><\/p>\n
Dessa helps the world\u2019s largest organizations reinvent what\u2019s possible with machine learning. Founded in 2016, the company has excelled at applying cutting-edge ML techniques to generate significant value in industries ranging from finance to telecommunications. Today, the company empowers teams to build and deploy enterprise-grade ML applications at scale with Foundations Atlas and Foundations Orbit. Laarn more at dessa.com<\/a>.<\/p>\nSource: Dessa<\/i><\/p>\nBackgrounder:<\/h3>\nHello, World! Introducing Atlas and Orbit, Our New Products For Engineering Real-World ML At Scale<\/h2>\n
Dessa has worked with some of the world\u2019s biggest companies, and Mohammed Ridwanul, Foundations\u2019 Product Manager, says there are common challenges holding teams back from delivering the full return on investment for their machine learning projects.<\/p>\n
In development \u2026<\/p>\n
\n- Siloed work<\/strong>: collaboration and transparency in modeling efforts was lacking, leading to duplicated efforts and a lack of knowledge transfer<\/li>\n
- Expensive GPU infrastructure<\/strong>: which went underutilized because engineers didn\u2019t have an easy way to run many experiments in parallel<\/li>\n
- Ad hoc experiment management:<\/strong> a lack of simplified ways to run, evaluate and share the 1000s of experiments necessary to build accurate models<\/li>\n
- Lack of reproducibility: <\/strong>reproducing experiment results was a rarely achieved feat, resulting in lots of time lost during development<\/li>\n<\/ul>\n
In production \u2026<\/p>\n
\n- Model degradation: <\/strong>models degrade with time because of business changes and concept drifts. No tools for tracking degradation meant inaccurate predictions and millions lost in revenue.<\/li>\n
- Difficult to validate data: <\/strong>low-quality tools to monitor inevitable changes to data due to changes in customer and business behavior.<\/li>\n<\/ul>\n
Without adequate tools to solve these challenges, it\u2019s nearly impossible for ML teams to deliver maximum ROI for their business.<\/p>\n
With years of successful ML deployments under our belts, we\u2019ve been on a mission to change this. That\u2019s why we\u2019re so excited to introduce Atlas and Orbit.<\/p>\n
Introducing Foundations Atlas (v1.0)<\/strong><\/h3>\nAtlas takes you from 0 to 100 fast, with tools that make it easy to run, manage, evaluate and deploy 1000s of experiments concurrently. Atlas runs on any infrastructure set-up (the cloud, on-premise GPU clusters, you name it) and was designed to make team collaboration a breeze.<\/p>\n
![](\"https:\/\/www.insurance-canada.ca\/wp-content\/uploads\/2019\/10\/Dessa-Foundations-Atlas-v1-300x150.jpeg\")
<\/a><\/p>\nKey features:<\/p>\n
\n- Async and parallel job execution:<\/strong> run thousands of jobs asynchronously and simultaneously to push your infrastructure to its limits.<\/li>\n
- Job dashboard<\/strong>: an access-controlled dashboard allows you to view and manage model development metrics, parameters and artifacts in real time, across your organization.<\/li>\n
- Preemptible GPUs: <\/strong>Atlas makes it easy to use preemptible GPUs, so you can save 8x on compute costs (vs. you know, regular GPUs).<\/li>\n
- Job-level labeling, tagging and note-keeping:<\/strong> features that allow you to manage your experiments like a highly organized data scientist.<\/li>\n
- Model packaging for easy reproducibility: <\/strong>Atlas wraps every job as a package with a unique UUID that contains everything needed to run your model, including the training code, trained model and all associated artifacts.<\/li>\n
- Slack and email integration:<\/strong> get real-time notifications about your jobs\u2019 status where you\u2019ll see them.<\/li>\n
- Method-level caching: <\/strong>save even more on compute costs with method-level caching.<\/li>\n
- SQL-based model registry integration:<\/strong> maintain accurate records of every model developed for your organization with an open API that imports every model package into your own SQL model registry database.<\/li>\n
- Single-click deploy: <\/strong>got a model you want to integrate with an app you have built? Or want to deploy it into production? Turn your models into REST-based web servers with a single step!<\/li>\n<\/ul>\n
Interested in trying Atlas? After reading this, we hope you can\u2019t wait to try out some of Atlas\u2019 features for yourself, which is why we\u2019re offering a free trial<\/strong> of the platform with minimal set-up required.<\/p>\nTry Atlas for free at atlas.dessa.com<\/a>.<\/p>\nAnnouncing Foundations Orbit<\/strong><\/h3>\nSad but true:<\/em> models degrade over time due to concept drift, data and business requirement changes.<\/p>\nOnce you deploy models into production, it\u2019s critical to carry out the same kind of due diligence and monitoring on your machine learning models as you would with regular old software. Orbit ensures that your models stay on the right path once they\u2019re in production, with a wide range of tools for automated performance tracking and maintenance management.<\/p>\n
![](\"https:\/\/www.insurance-canada.ca\/wp-content\/uploads\/2019\/10\/Dessa-Foundations-Orbit-300x147.jpg\")
<\/a><\/p>\nTo sign up for a demo or learn how to implement Orbit for your enterprise, please get in touch with us at [email protected]<\/span><\/a>.<\/p>\n\nThere is nothing that brings us more joy at Dessa than removing barriers to ML creation \u2014 whether you\u2019re building a model for fun, trying to solve global issues, or to save your company millions of dollars (or all of the above), we\u2019d love to help. Reach out to us at [email protected]<\/a> or talk to me personally on our Intercom at www.dessa.com, to see how we can help you make things happen
Mohammed Ridwanul, Product Manager \u2013 Foundations Atlas & Orbit<\/cite><\/p>\n<\/blockquote>\nAbout Dessa<\/b><\/p>\n
Dessa helps the world\u2019s largest organizations reinvent what\u2019s possible with machine learning. Founded in 2016, the company has excelled at applying cutting-edge ML techniques to generate significant value in industries ranging from finance to telecommunications. Today, the company empowers teams to build and deploy enterprise-grade ML applications at scale with Foundations Atlas and Foundations Orbit. Laarn more at dessa.com<\/a>.<\/p>\nSource: Dessa <\/i><\/p>\n
Tags: Artificial Intelligence (AI)<\/a>, Dessa<\/a>, launch<\/a>, Machine Learning (ML)<\/a> Read the original article at Insurance-Canada.ca <\/a><\/p>\n","protected":false},"excerpt":{"rendered":"\u2018Machine Learning Superpowers\u2019: Announcing Atlas and Orbit \u2013 products for teams that want to rapidly scale their ML efforts Dessa has distilled its experience engineering ML for the world\u2019s largest companies into powerful ML...<\/p>\n","protected":false},"author":578,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[],"tags":[],"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/posts\/17251"}],"collection":[{"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/users\/578"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/comments?post=17251"}],"version-history":[{"count":0,"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/posts\/17251\/revisions"}],"wp:attachment":[{"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/media?parent=17251"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/categories?post=17251"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/tags?post=17251"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
- \n
- Atlas:<\/strong> Tools for rapidly developing ML at scale. Enables ML engineers to manage 1000s of experiments concurrently, save 8x on compute costs and deploy models with a single click<\/li>\n
- Orbit:<\/strong> Tools for managing ML models in production. Enables ML engineers to overcome costly production challenges, with features for model performance monitoring, maintenance, automatic training and A\/B testing<\/li>\n<\/ul>\n
\u201cAtlas transforms the time-consuming and frustrating process of managing experiments into something that\u2019s easy and even enjoyable, also allowing for insights you might not have otherwise,\u201d said Alex Krizhevsky, Principal ML Architect at Dessa.<\/p>\n
The design of both products was informed by the company\u2019s experience deploying ML for the world\u2019s largest organizations, including Fortune 500 companies. Stephen Piron, the company\u2019s Co-CEO remarks that \u201cbefore Foundations, helping these companies build and deploy their first AI systems was never easy. It often felt like we had to drag these systems kicking and screaming into production. But everything we\u2019ve learned has been baked into Atlas and Orbit. We hope offering these tools to the world will be truly game-changing.\u201d<\/p>\n
Teams wanting to test the products themselves can register for a free trial of Atlas<\/a>, also featuring Google Cloud credits and a text-generation tutorial. Demos for Atlas and Orbit are also available.<\/p>\n
About Dessa<\/b><\/p>\n
Dessa helps the world\u2019s largest organizations reinvent what\u2019s possible with machine learning. Founded in 2016, the company has excelled at applying cutting-edge ML techniques to generate significant value in industries ranging from finance to telecommunications. Today, the company empowers teams to build and deploy enterprise-grade ML applications at scale with Foundations Atlas and Foundations Orbit. Laarn more at dessa.com<\/a>.<\/p>\n
Source: Dessa<\/i><\/p>\n
Backgrounder:<\/h3>\n
Hello, World! Introducing Atlas and Orbit, Our New Products For Engineering Real-World ML At Scale<\/h2>\n
Dessa has worked with some of the world\u2019s biggest companies, and Mohammed Ridwanul, Foundations\u2019 Product Manager, says there are common challenges holding teams back from delivering the full return on investment for their machine learning projects.<\/p>\n
In development \u2026<\/p>\n
- \n
- Siloed work<\/strong>: collaboration and transparency in modeling efforts was lacking, leading to duplicated efforts and a lack of knowledge transfer<\/li>\n
- Expensive GPU infrastructure<\/strong>: which went underutilized because engineers didn\u2019t have an easy way to run many experiments in parallel<\/li>\n
- Ad hoc experiment management:<\/strong> a lack of simplified ways to run, evaluate and share the 1000s of experiments necessary to build accurate models<\/li>\n
- Lack of reproducibility: <\/strong>reproducing experiment results was a rarely achieved feat, resulting in lots of time lost during development<\/li>\n<\/ul>\n
In production \u2026<\/p>\n
- \n
- Model degradation: <\/strong>models degrade with time because of business changes and concept drifts. No tools for tracking degradation meant inaccurate predictions and millions lost in revenue.<\/li>\n
- Difficult to validate data: <\/strong>low-quality tools to monitor inevitable changes to data due to changes in customer and business behavior.<\/li>\n<\/ul>\n
Without adequate tools to solve these challenges, it\u2019s nearly impossible for ML teams to deliver maximum ROI for their business.<\/p>\n
With years of successful ML deployments under our belts, we\u2019ve been on a mission to change this. That\u2019s why we\u2019re so excited to introduce Atlas and Orbit.<\/p>\n
Introducing Foundations Atlas (v1.0)<\/strong><\/h3>\n
Atlas takes you from 0 to 100 fast, with tools that make it easy to run, manage, evaluate and deploy 1000s of experiments concurrently. Atlas runs on any infrastructure set-up (the cloud, on-premise GPU clusters, you name it) and was designed to make team collaboration a breeze.<\/p>\n
<\/a><\/p>\nKey features:<\/p>\n
- \n
- Async and parallel job execution:<\/strong> run thousands of jobs asynchronously and simultaneously to push your infrastructure to its limits.<\/li>\n
- Job dashboard<\/strong>: an access-controlled dashboard allows you to view and manage model development metrics, parameters and artifacts in real time, across your organization.<\/li>\n
- Preemptible GPUs: <\/strong>Atlas makes it easy to use preemptible GPUs, so you can save 8x on compute costs (vs. you know, regular GPUs).<\/li>\n
- Job-level labeling, tagging and note-keeping:<\/strong> features that allow you to manage your experiments like a highly organized data scientist.<\/li>\n
- Model packaging for easy reproducibility: <\/strong>Atlas wraps every job as a package with a unique UUID that contains everything needed to run your model, including the training code, trained model and all associated artifacts.<\/li>\n
- Slack and email integration:<\/strong> get real-time notifications about your jobs\u2019 status where you\u2019ll see them.<\/li>\n
- Method-level caching: <\/strong>save even more on compute costs with method-level caching.<\/li>\n
- SQL-based model registry integration:<\/strong> maintain accurate records of every model developed for your organization with an open API that imports every model package into your own SQL model registry database.<\/li>\n
- Single-click deploy: <\/strong>got a model you want to integrate with an app you have built? Or want to deploy it into production? Turn your models into REST-based web servers with a single step!<\/li>\n<\/ul>\n
Interested in trying Atlas? After reading this, we hope you can\u2019t wait to try out some of Atlas\u2019 features for yourself, which is why we\u2019re offering a free trial<\/strong> of the platform with minimal set-up required.<\/p>\n
Try Atlas for free at atlas.dessa.com<\/a>.<\/p>\n
Announcing Foundations Orbit<\/strong><\/h3>\n
Sad but true:<\/em> models degrade over time due to concept drift, data and business requirement changes.<\/p>\n
Once you deploy models into production, it\u2019s critical to carry out the same kind of due diligence and monitoring on your machine learning models as you would with regular old software. Orbit ensures that your models stay on the right path once they\u2019re in production, with a wide range of tools for automated performance tracking and maintenance management.<\/p>\n
<\/a><\/p>\nTo sign up for a demo or learn how to implement Orbit for your enterprise, please get in touch with us at [email protected]<\/span><\/a>.<\/p>\n
\n
There is nothing that brings us more joy at Dessa than removing barriers to ML creation \u2014 whether you\u2019re building a model for fun, trying to solve global issues, or to save your company millions of dollars (or all of the above), we\u2019d love to help. Reach out to us at [email protected]<\/a> or talk to me personally on our Intercom at www.dessa.com, to see how we can help you make things happen
Mohammed Ridwanul, Product Manager \u2013 Foundations Atlas & Orbit<\/cite><\/p>\n<\/blockquote>\nAbout Dessa<\/b><\/p>\n
Dessa helps the world\u2019s largest organizations reinvent what\u2019s possible with machine learning. Founded in 2016, the company has excelled at applying cutting-edge ML techniques to generate significant value in industries ranging from finance to telecommunications. Today, the company empowers teams to build and deploy enterprise-grade ML applications at scale with Foundations Atlas and Foundations Orbit. Laarn more at dessa.com<\/a>.<\/p>\n
Source: Dessa <\/i><\/p>\n
Tags: Artificial Intelligence (AI)<\/a>, Dessa<\/a>, launch<\/a>, Machine Learning (ML)<\/a> Read the original article at Insurance-Canada.ca <\/a><\/p>\n","protected":false},"excerpt":{"rendered":"
\u2018Machine Learning Superpowers\u2019: Announcing Atlas and Orbit \u2013 products for teams that want to rapidly scale their ML efforts Dessa has distilled its experience engineering ML for the world\u2019s largest companies into powerful ML...<\/p>\n","protected":false},"author":578,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[],"tags":[],"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/posts\/17251"}],"collection":[{"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/users\/578"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/comments?post=17251"}],"version-history":[{"count":0,"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/posts\/17251\/revisions"}],"wp:attachment":[{"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/media?parent=17251"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/categories?post=17251"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.lifeinsurance-orleans.ca\/index.php\/wp-json\/wp\/v2\/tags?post=17251"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
- Job dashboard<\/strong>: an access-controlled dashboard allows you to view and manage model development metrics, parameters and artifacts in real time, across your organization.<\/li>\n
- Difficult to validate data: <\/strong>low-quality tools to monitor inevitable changes to data due to changes in customer and business behavior.<\/li>\n<\/ul>\n
- Expensive GPU infrastructure<\/strong>: which went underutilized because engineers didn\u2019t have an easy way to run many experiments in parallel<\/li>\n
- Orbit:<\/strong> Tools for managing ML models in production. Enables ML engineers to overcome costly production challenges, with features for model performance monitoring, maintenance, automatic training and A\/B testing<\/li>\n<\/ul>\n