The Maslow’s hierarchy your data science team needs
Domino Data Lab was announced as a leader for the second year in a row in the recently released “Forrester Wave™: Notebook-based Predictive Analytics and Machine Learning (PAML), Q3 2020” analyst report. True to our data science roots, we’ve built a Maslow’s hierarchy of data science team needs.
By David Bloch
Domino Data Lab was announced as a leader for the second year in a row in the recently released “Forrester Wave™: Notebook-based Predictive Analytics and Machine Learning (PAML), Q3 2020” analyst report. The report highlighted Domino’s capability to “support the diversity of [machine learning] ML options with repeatability, discipline, and governance” while also mentioning the collaboration and governance capabilities that we believe differentiate Domino as a market leader.
We believe this report is validation for our product vision and strategy. At Domino, we work with some of the most sophisticated data science organizations in the Fortune 100, and as a result, have developed an informed view on the critical components that data science teams need for success with an advanced data science platform. While scalable compute, access to code libraries, and repository management are essential, they are merely a foundation layer and table stakes for data science platforms.
True to our data science roots, we’ve built a Maslow’s hierarchy of data science team needs. Here, we plotted the elements that we believe data science teams need to satisfy, and offer an approach to help them drive discipline, process, rapid iteration, and scale via consistent solution delivery to their stakeholders.
The critical aspects to a hierarchy of needs pyramid start in ascending order include:
- Scalable Compute across CPU, GPU, and APU at the foundation level and access to all data repositories
- Centralized Tooling that is open, extensible, and flexible (required by researchers) comes next
- User Access Control and Security to provide access, role specialization, and authentication across tooling.
- Versioning and version control across all data science assets
- End to end orchestration to enable reproducible experimentation, task scheduling, and connectivity
- A dynamic, collaborative user interface that supports team-based collaboration and problem-solving
- Knowledge Management and Governance solutions that make it easy to track all decisions, outputs, and outcomes in a single place and monitor models through their lifecycle.
- The ability to deploy data science products into the hands of end-users in a simple, repeatable fashion.
Domino is a platform built by data scientists for data science teams. Our primary focus is on fulfilling data science teams’ needs to help them produce data science products that drive impact to the business, not to sell your team more cloud storage like some of our competitors. These products are used by business stakeholders to transform the way their business operates and create business value.
We are committed to remaining fully open and extensible, bringing together workloads in traditional enterprise analytics suites such as SAS and MATLAB with cutting edge open-source technologies like Spark, Python, and R.
We look end-to-end across the entire life cycle of model development and delivery, with innovative solutions that help deliver the “last-mile” of taking a machine learning model from lines of code in a repository into a deployed business solution.
This focus on delivering data science products, rather than just data science projects, encourages teams to foster a culture of collaboration and communication among data scientists and with their business counterparts from across the organization.
By focusing on delivering data science products into the business, the Domino platform helps foster healthy stakeholder relationships and communications via a full suite of model monitoring, management, and collaboration features. We believe we see the world differently from most data science platform vendors who solely focus on delivering code and compute.
We do this by offering various ways to deploy models into production backed by a reproducibility and collaboration engine that lets you see all of the assumptions, decisions, and actions that data science teams take when choosing the right method for problem-solving.
This commitment means we remain focused on solving problems that enterprise organizations face when bringing the promise of AI/ML toolsets to critical business solutions—that can help them differentiate themselves and build sustainable competitive advantage in their respective industries.
Today, we are pleased to support data science at scale in over 20% of the Fortune 100. We have a unique vision for where notebook-based predictive analytics and machine learning platforms need to go next, and we look forward to continuing to realize that vision with our customers. We believe that Forrester feels as we do by giving Domino the the highest scores possible in the collaboration and platform infrastructure criteria, among the top scores in the ModelOps criterion, and among the second highest scores in the Strategy category.
Get your complimentary copy of this critical research report today to see the rest of our scores and how the competition performed.