DataRobot

 

“Empowering the human heroes of the intelligence revolution

Founded: 2012, Boston, MA, USA

Category: Artificial Intelligence/Enterprise Software, Information Technology, Machine Learning, Predictive Analytics, Robotics, SaaS

Primary office: Boston, MA, USA

Core technical team: Boston, MA, USA

Status: Private

Employees: 1001-5000

Amount raised: USD$431 million (6 rounds – Sept 2019)

OVERVIEW

  • Industry agnostic AI enterprise platform applicable across sectors and roles – forms the AI technology base for companies adopting AI to create value
  • DataRobot platform allows customers to prepare their data, create and validate machine learning models, including time series models, and deploy and monitor those models in a single solution
  • Distinct competences – automated machine learning
  • Marriage of robotic process automation and machine learning
  • Invented the automated machine learning category/Patents in predictive analytics

PERFORMANCE METRICS

  • Valuation: USD $1 billion (2019)
  • Estimated Revenue: N/A
  • Triple digit growth since 2015
  • 3,000+ companies helped
  • One third of Fortune 50 are clients

 

ACHIEVEMENTS

  • Almost 2 billion models built
  • At least 20 patents

Sells

  • DataRobot enterprise AI platform (allows customers to prepare their data, create and validate machine learning models, including time series models, and deploy and monitor those models in a single solution). The platform consists of:
    • Paxata data prep – visually and interactively explore, combine, and shape diverse datasets into data ready for machine learning and AI applications at enterprise scale
    • Automated machine learning – automate the creation of advanced machine learning models that incorporate our world-class data science expertise
    • Automated time series – automate the development of sophisticated time series models that predict the future value of a data series based on its history and trend
    • MLOps – deliver the capabilities that data science and IT Ops teams need to work together to deploy, monitor, and manage machine learning models in production
  • Platform available in the cloud, on premise, or as fully managed AI service
  • ROI enablement services in support of platform
  • Licenses are on a per user basis

Channels

  • Partnership programs – technology alliances (innovative integrations), value added resellers (scalable partnering environment), solution partners, system integrators and consulting firms
  • Academic licenses and non-profit licenses (“use DataRobot for good today”; up to 30 days)
  • DataRobot community – messaging platform (discussions)
  • DataRobot University
  • Newsletter – data science news, insights, tutorials, etc.
  • Wiki
  • Webinars
  • Podcasts

Competencies

  • Marrying robotic process automation with machine learning; according to company literature, they invented the automated machine learning category
  • Significantly simplifies the use of AI/ML through high degree of automation – democratizing AI
  • Pure AI play across sectors with a large number of proven use cases (by sector and by role)
  • Visualizations of models using commodity hardware
  • Simplified process – connect to data, drag and drop, automatically evaluate 100s of models, monitor and manage all deployed models
  • Runs competitions between AI algorithms to determine which are best for the task at hand
  • Partnership ecosystem to help customers become AI-driven enterprises through the adoption of AI capabilities

Distinct AI Features

Type

  • Large collection of open source AI/ML algorithms
  • Proprietary predictive analytics algorithms (patented)
  • Smart hyperparameter tuning and ensemble models (combine strengths of several algorithms and balance out weaknesses of others)

 

  AI use

  • DataRobot’s primary use of AI is to provide an AI/ML platform to enterprises across business sectors.
  • Supplier of AI competencies and packages such in a manner that facilitates adoption.
  • Developed proprietary algorithms in predictive analytics (various patents).

Rate of return on customer’s investment to make AI work

Immediate:

  • Accessible AI/ML functionality to a broader community (than pure data scientist); Numerous use cases to demonstrate facility

Long term:

  • Customer’s organization is an AI/ML-based enterprise with revolutionary new approaches to making its most important decisions

Databases

  • Databases are generally client specific

Quantum Computing

  • No explicit mention of quantum computing in DataRobot collateral. Generically, quantum computing is expected to enable artificial intelligence by, for example, handling huge amounts of data, building better models and more accurate algorithms, and using multiple datasets (integrating datasets more quickly)

Resources

Assets

  • Open source ML algorithms
  • DataRobot platform
  • AI Catalog – helps clients find, understand and use the data they need for their projects in a governed AI platform. AI Catalog is the heart of the DataRobot Platform in enabling searching, collaboration, and sharing of assets for AI projects
  • AI-native strategic success team (turn data into value)
  • 1000+ total years data science experience on customer-facing data science team
  • 4 million person hours of engineering innovation to build platform
  • Diverse partnership program
  • Platform is architected in such a manner that it can support service providers (value added resellers, solution partners, system integrators and consulting firms).
  • R&D teams

Processes

  • Defined the 10 steps of automated machine learning: data identification, data preparation, feature engineering, algorithm diversity, algorithm selection, training and tuning, head-to-head model competitions, human friendly insights, easy deployment, and model monitoring and management
  • Platform available in the cloud, on premise, or as fully managed AI service
  • Supports technology alliances in which complementary innovative technologies are integrated with DataRobot’s offerings.
  • 30 end-to-end guides that provide a framework and best practices on how to implement AI within an organization
  • Ongoing development of use cases across industries including airlines, automotive, financial services, gaming, healthcare, higher education, industry agnostic, insurance, manufacturing, media, nonprofit, retail, sports, technology and telecom

Priorities

  • Invented the automated machine learning category and intend to keep reinventing the category
  • Delivering powerful AI and ML solutions that are relevant and accessible to all – democratizing AI

References

  1. DataRobot website. http://datarobot.com
  2. Evaluation: Driverless AI vs. DataRobot. https://wyzoo.com/blog/driverlessai-vs-datarobot.html
  3. Patents assigned to DataRobot, Inc. https://patents.justia.com/assignee/datarobot-inc
  4. A marriage of robotic process automation and machine learning. https://www.forbes.com/sites/tomdavenport/2019/06/06/a-marriage-of-robotic-process-automation-and-machine-learning/#1932bde07947
  5. Techcrunch. Enterprise companies find MLOps critical for reliability and performance. https://techcrunch.com/2020/05/06/enterprise-companies-find-mlops-critical-for-reliability-and-performance/

Contributors

  • Dan Craigen