Dotscience Looks To Simplify MLOps and Speed Delivery of AI Projects

Dotscience is making it easier to deploy and monitor machine learning models on Kubernetes clusters. IDN looks at Dotscience Deploy and Monitor, which offers several benefits to promote MLOps, such as helping data science and machine learning teams promote more sharing and easier provisioning.

Tags: AI, API, cloud, deploy, DevOps, Docker, Dotscience, Kubernetes, machine learning, MLOps,

Dotscience is rolling out features to make it easier to deploy and monitor machine learning models on Kubernetes clusters. 


"By simplifying deployment and monitoring of models, Dotscience is making MLOps accessible to every data scientist -- without forcing them to set up and configure complex and powerful tools like Kubernetes, Prometheus, and Grafana from scratch," said Dotscience CEO and co-founder Luke Marsden in a statement.


Dotscience's DevOps for ML platform was designed with features to simplify, accelerate and control all stages of the AI model lifecycle.


"Our push to simplify deployment and monitoring of AI/ML is based on the market insight that many businesses are still struggling with deploying their ML models, blocking any business value from AI/ML initiatives," he said in a statement. 


The platform's Dotscience Deploy and Monitor feature offers several benefits to data science and machine learning teams to promote more sharing and easy provisioning, he added. 


Dotscience enables data science and ML teams to own and control the entire model development and operations process, from data ingestion, through training and testing, to deploying straight into a Kubernetes cluster, and monitoring that model in production to understand its behavior as new data flows in. This complete end-to-end approach overcomes the limits of handling only portions of the ML development and operations process, according to Marsden. It also avoids the need for add-on integration for such end-to-end functionality, he added. 


To make Kubernetes simple and accessible to data scientists, Dotscience lets data science and ML teams use a single click (API call or command) to perform a rich set of crucial tasks through the entire lifecycle. The latest update enables teams to ingest data, perform data engineering, train and test models, pre-test within CI pipelines and even deploy them to production. 


Dotscience also simplifies the deploying of ML models to Kubernetes. Deployment is also easier for users, as they can deploy their models in three main ways:

  • UI deploymentsAfter defining parameters in the UI, users can deploy straight from within the Dotscience Hub interface
  • CLI style: The Dotscience CLI tool 'ds' can be used to deploy an ML model using command line parameters to define the exact details
  • From the Python library: Deploy directly from the python library with ds.publish (deploy=True), which also automatically sets up a statistical monitoring dashboard. 

Dotscience also offers flexibility across hybrid environments. As Marsden explained it: 


"By enabling hybrid and multi-cloud scenarios, where training happens on-prem where the data is, and the deployment to production and inference happens in the cloud where Kubernetes is easy to set up, we enable flexible use of on-prem infrastructure along with easy access to harness the power of Kubernetes in the cloud." 


This approach aims to lets teams design models that can be statistically monitored across all phases of design, iteration and go live.


In Dotscience's latest updated, users can:

  • Handle both building the ML model into a Docker image and deploying it to a Kubernetes cluster
  • Gain optionality for handling CI/CD responsibilities. Teams can use existing infrastructure or employ lightweight built-ins
  • Track the deployment of the ML model back to the provenance of the model (including the data used to train the model) This approach lets users maintain accountability across the entire lifecycle
  • Easy access to in-depth metrics and dashboards.

On this last note, Marsden noted the connection between the Dotscience architecture and end-to-end visibility. 


With its latest extension, the Dotscience end-to-end support provides both metrics and context. This lets users monitor the health of their ML models at every stage of the lifecycle.


"Monitoring models in ML-specific ways is not obvious to software-focused DevOps teams... Often there's a disconnect between the type of monitoring performed by operations teams, such as error rates and request latency, and the type of monitoring that machine learning teams need to do on their models when deployed to production, such as looking at the statistical distribution of predicted categories," Marsden noted. 


In real-life situations, this end-to-end approach efficiently puts all team members on the same page by providing insight into what Marsden called "context-specific monitoring information about their model." Having this more in-depth insight "better positions them to understand why an error occurred and respond to it, rather than putting this onus on a central operations team," he added.