Agent Workforce Platform for AWS

DataRobot provides an analytic platform that enables me to rapidly prototype and model, as well as deploy applications and integrations between models and application stacks, and third party apps. It's a wonderful platform.

A specific example of how I've used DataRobot for a project is that we've used it in many use cases. We've looked at it from a financial department perspective and managing accounts receivable and accounts payable and time to payment, to HR related activities where we're monitoring employee churn, attribution, and trying to understand how quickly the churn is or what the cost is for HR related hiring. We looked at it from a business perspective and how to model some of our processes and automate them. The platform provided the ability to integrate into these data sources and provides the ability to work with Docker containers, work with cloud providers, and move the outputs to wherever they need to be. A long list of business use cases have been successfully delivered because of the platform.

My main use cases for DataRobot are the ability to rapidly model data to understand directionally where and what would be needed, which is where DataRobot is ahead of the class. The ability to begin a project, model it, and get in front of stakeholders within hours is something I've actually performed in a meeting as a live demonstration. It's a very solid platform, especially for analytics and generative AI and needing to have guardrails around production grade models.

DataRobot has positively impacted my organization by first providing the ability for us to rapidly go through numerous use cases and assess ROI on these use cases such that as an organization, we could reprioritize these projects based upon impact to the organization. It's an eye-opening results that the system can provide.

A specific outcome I've experienced with DataRobot is the time saved, which would be an interesting discussion. In the past, data scientists would build a model and be able to get their predictions through the model but would have a very difficult time getting this model into production and up through the different cloud vendors and all of the hardware. It was a compatibility issue. DataRobot knocked down those walls so that production grade models can be delivered within minutes, no longer two week sprints waiting for front-end teams to deploy your model. You can deploy it within minutes, and that's the time saving.

The best features DataRobot offers include integration with any cloud platform, Kubernetes, Docker containers, APIs, real time APIs, and batch APIs. I'm able to source data in near real time, source data in batch. I'm able to output data in near real time or output results from a model in batch. This platform is the spoke and hub, which is the center of a spoke and hub model.

Out of all those features, the one I find myself using most often is the ability to model a project's data, along with the ability to deploy that model with either REST APIs for batch or for near real time activity. The whole process of building the model and then deploying it and then making it available for people to use with a governed dashboard, which is extremely important. The ability to monitor the predictions and the data to understand if predictions are drifting or to understand if incoming newer data is drifting is a complete ecosystem that helps you deliver models for your organization.

DataRobot is ahead of the curve when it comes to generative AI and these platforms and explainable AI. Organizations tend to rely too heavily on cloud providers and solution architects from those cloud providers to create a large architectural engineering effort, when in fact, by simply leveraging DataRobot, your platform is in place. Your models can be deployed as soon as they're done training, and that's done within minutes without any heavy organizational lift. You can build and deploy models with endpoints within minutes with very little organizational heavy lift or requirements on other teams to get your model into production.

Regarding DataRobot's AI capabilities, I think the governance and security are already answering the questions that other enterprises are struggling with today. DataRobot is backing into it. DataRobot already has those integration, architectural integration patterns in place for security, for access, and for permissions, and for many other security measures that you would want to manage and maintain. DataRobot out of the box already supplies that functionality.

Regarding DataRobot's AI capabilities, I think its accuracy and reliability of output are outstanding. The ability to monitor and manage accuracy, the ability to understand data drift and prediction drift are key features. The tools are very proactive in monitoring its own behavior. When you do that, you can see that it already has a dashboard built for this type of information, and it can go even further if you customize it to your organization's needs. But right out of the box, you have a governance and compliance dashboard with all of the analytic assets right on day one.

DataRobot can be improved by continuing with the hub and spoke model, which will need to improve integration patterns, leveraging different programming languages. I would highly recommend that it starts to work with a programming language GoLang so that it can provide secure connections and secured integrations with all these other platforms that it already can work with today.

I started with DataRobot in 2019.

DataRobot is very stable.

DataRobot can scale at the level of any cloud provider with its integration patterns. Compute, storage, memory, anything to do with models or deployments of apps on Docker containers on Kubernetes servers, DataRobot is well positioned to help any and all organizations.

DataRobot can scale as far as is as deep and as wide as a cloud provider will give it access to.

The customer support I experience with DataRobot is excellent. They provide first-class documentation, first-class engineering teams, and first-class business representatives. They are a strong organization.

I would rate the customer support a 10. I've had such positive interactions on checking up on certain things or getting other things clarified, and I'm a big supporter.

We did not use a different solution before DataRobot. We reviewed five or six different vendors, and DataRobot was the platform that enabled us to do the entire analytical process workflow, start to finish, within the platform without needing any additional support. Whereas other vendors did have positives in certain parts of a machine learning workflow, none of the other vendors at the time we did the analysis could complete the entire process end to end the way DataRobot can.

I have deployed DataRobot in two manners. I have worked with it on-prem, where I did the complete install with very talented subject matter experts and an engineering team at DataRobot. We did install it on our own hardware. Fast forward a couple of years, and we moved the solution to a vendor-based managed solution, so a SaaS solution. The vendor took care of managing the platform, and that enabled us in data science to focus on the data science. The tool is the same, depending on the requirements of the environment, which path you take. I have worked with two different installations of the platform, and both have been awesome.

I have seen a return on investment, and I can share that money saved and time are always interconnected. Time would be the massive reduction in cost and deliverable by reducing the need for external teams for deployment of analytical assets or applications.

My experience with pricing, setup cost, and licensing is that I was part of the team that negotiated contracts and reviewed these proposals for setup costs. I do have firsthand experience in working with DataRobot's organization.

I find that DataRobot's pricing and licensing model is very much a hub and spoke model. The price of the platform is the price of the platform. What people often don't realize is because you're managing your cloud infrastructure build-out through the platform, DataRobot reduces third party costs on clouds because the platform builds in automatic shutdowns and management tools for things that you are working on with cloud providers. It is a cost-saving tool in cloud computing today.

Before choosing DataRobot, we evaluated three or four other vendors. They were all incomplete at the time. The only vendor that provided a complete analytical workflow was DataRobot and the platform that it provides.

I would rate DataRobot a 10 on a scale of 1 to 10.

I choose 10 because DataRobot stood out to me because the platform provided the answers to questions that are being asked by many organizations today. These questions include how do I manage my AI environment, how do I control my data access, how do I deploy apps, and how do I work with my models. They have the complete workflow automated. This is the platform you start with. This is the platform that becomes the hub and spoke model for organizations to rapidly move in today's environment with predictive and generative AI in the current business workflow.

The cloud provider I use for the SaaS or managed solution is that DataRobot is cloud agnostic. It can work with any cloud vendor. The requirements of an organization would shape which cloud provider that you would integrate with. DataRobot integrates with all of those on day one, right out of the box.

If you're looking into evaluating DataRobot, you need to understand the entire workflow for machine learning. When you see the entire workflow and you're able to do it in the platform, it's lightning in a bottle. Where people, data scientists such as myself, may get tripped up is that they're used to working one way, and the platform has a slightly different approach. The first couple of times getting used to finding where everything is, don't let that impact your decision on the functionality. It's just learning where the button is, not that the functionality isn't there. It's an amazing platform.