By Victor Ghadban

You already know that implementing an enterprise-ready AI enablement platform is key to executing your organization’s AI and machine learning initiatives. But can software so complex really be easy to implement? How can you avoid disruptions to your business? What can you do to prepare?

After spending the last 20 years working closely with customers in the software analytics space, I have learned a thing or two about how to run a smooth implementation process. Your organization can avoid operational disruptions, time delays, and cost overruns as long as you break the process down into five simple questions: Why, What, Who, Where, and How.

Why do you need an AI enablement platform?

Chances are you already have data scientists coding in Python and using the open source Anaconda Distribution to support their development. (I can only assume they’re using Anaconda Distribution…unless they actually like downloading numerous packages from unknown sites and spending countless hours getting new packages and dependencies into place before they can start writing code?) With the Anaconda Distribution, they can work on their laptops using a Notebook to develop cool new AI models in Python, with Anaconda Distribution doing all of the heavy lifting.

But as your company grows, you start to lose oversight of what people are developing. Code is scattered all over the place, with copies flying around as users commit and pull from Git, and, worst case—copies of sensitive PII data residing on individual laptops.

Obviously, a laptop can only get you so far. So what happens when you need to scale up your AI and do it quickly?

Successful AI in the enterprise involves large volumes of data, and to process that data you need a tool that can provide deep learning libraries to develop AI models, work with GPUs for speed, and give you the ability to run through heavy computational code, visualize, test, and deploy.

You need an AI enablement platform.

What platform should you choose?

When companies want to join the world of AI, they don’t need to look any further than our AI enablement platform, Anaconda Enterprise, to fulfill all of their needs. Anaconda Enterprise is a single platform that allows data scientists to develop, test, visualize, and deploy their code all in one place.

Our customers love Anaconda Enterprise because it automates your organization’s AI pipelines from laptops to training clusters to production clusters with ease, and supports your organization no matter the size, scaling from a single user to thousands, running on one laptop to thousands of machines. You can deploy anything you create with Anaconda Distribution into a production-like environment with just one click.

Who is needed to implement your platform?

When installing and configuring such a complex enterprise platform, it’s crucial to have the right team members in place.

For the installation, Anaconda carried the heavy lifting by making it a seamless single installer for everything you’ll require. There is no need for additional software download or installs, no need for Java, no need for C compilers or any other type of supportive software. For Anaconda Enterprise, the cleaner the server is, the easier it is to install.

But the configuration will take a little more work, and this is where assembling the right people from the right teams comes in. For security, Anaconda can plug into LDAP, AD, Kerberos, and SAML for Authentication, but if you have an LDAP/AD admin dedicated to the project to give you the right parameters and credentials to connect and federate the platform, you can get your users up and running in no time.

Just as Anaconda can plug into many security utilities, it can plug into a range of data sources, including Hadoop, NoSQL, SQL, Cloud Data Stores, and Flat file. Chances are the respective administrators locked down these sources pretty well. If you have the appropriate admin involved in helping set up the connectivity and user access to these sources, the quicker you will have your users running large queries and training extensive deep learning models.

It takes a total team effort from both sides, but when you have the right people available who can provide the parameters to connect to external sources, the installation and configuration can be a pretty painless process.

Where is Anaconda Enterprise going to live?

Anaconda Enterprise is a fully contained AI and machine learning platform that enables easy development, deployment, and governance using a cloud native architecture to support. If you’ve checked out the cloud native webinar presented by Anaconda SVP of Products & Marketing Mathew Lodge, you’ll understand that being cloud native allows Anaconda Enterprise to be installed anywhere—whether it’s in a popular cloud provider, a private cloud, or on-premise on bare metal or VMs. This flexibility allows you to choose the appropriate environment while not being fully locked down, and gives you the ability to move to the cloud or even back to on-premise if you so choose.

Once you’ve chosen the environment, where exactly in that environment should Anaconda Enterprise live? For complex AI development, data retrieval and processing will be fundamental. You will want to locate Anaconda Enterprise within a subnet close to or directly near the data you will be retrieving. This can be a Hadoop ecosystem, database, or file store. You can also bring the data directly into Anaconda Enterprise with a centralized Network File System to share data among numerous projects and people.

How is this all done?

It starts with an Anaconda representative meeting with your company to determine if Anaconda Enterprise is the right solution for you. One of our sales engineers will meet with your team to understand the problems you’re trying to solve and to present a demonstration of the product to help answer any questions you might have.

An Anaconda solutions architect will meet with your team to understand the use cases, growth patterns, and user activity to help size a cluster to fit your needs. Once we’ve defined a clear statement of work, assigned a dedicated project manager to lead the way, and assembled the implementation team, Anaconda can help your company move forward and enter the exciting world of AI.

Interested in learning more?

Ed. note: You can hear Victor talk more about his experiences with the implementation process in our on-demand webinar, Best Practices for Implementing an Enterprise AI Platform!
Best Practices for Implementing an Enterprise AI Platform