MLOps

Machine Learning Operations
Join this virtual event and interactive discussion with industry experts about operationalizing machine learning models in the enterprise. Learn how HPE (with solutions from BlueData) is helping address the challenges of deploying and managing machine learning models at scale.
jameskobielus
In just moment, we'll post question 7 (of 8).
jameskobielus
Less than 15 minutes to go in today's #MLops CrowdChat. It's good to see all the contributions.
Storage Godfather (HPEStorageGuy)
Here's another new video to check out after the chat: https://www.youtube.com/watch?v=c97wVQJo-cA
https://www.youtube.com/embed/c97wVQJo-cA
Using HPE ML Ops to drive success with Artificial Intelligence
Using HPE ML Ops to drive success with Artificial Intelligence
Victor Ghadban, Field CTO at HPE (BlueData) talks about how enterprises can drive success with AI using HPE ML Ops. HPE ML Ops is a secure, highly scalable s...
jameskobielus
Q6: How is the ML model lifecycle different than DevOps? What aspects of DevOps can be applied to ML models? What aspects are unique to ML Ops? https://www.crowdchat.net/s/35vit
https://www.crowdchat.net/s/35vit

Patrick Osborne
In many ways they extremely adjacent
Mike Leone
While DevOps is more about marrying the developers and IT operations folks, ML Ops builds on DevOps and incorporates some additional personas, such as data engineers and data scientists.
Frederic Van Haren
MLops is building on top of the DevOPS concepts. However, MLops focusses on the ML lifecycle which is a pipeline that relies on continuous feedback and improvement.
Patrick Osborne
In order to scale models and the implementation of models and to quickly train, infer and re-run models...one has to do this with a DevOps framework
Mike Leone
For ML Ops specifically, it's heavily centered on providing intelligence derived from data to a specific business objective, goal, or use case, and enabling developers to consume models and utilize the insights for rapid innovation.
Victor Ghadban
Training a model and reiterating through the cycle until you achieve the acceptance criteria of the model. Devops doesn’t have this requirement or iterative step.
Abdul Matheen Raza
ML Lifecycle is very iterative with a heavy dependency on data. MLOps leverages DevOps processes, however the iterative nature of the ML Lifecycle and the dependency on data should be taken into account. Changes in underlying data will drive retraining decisions
Storage Godfather (HPEStorageGuy)
This video talks about operationalizing #MLOps
https://www.youtube.com/watch?v=8p1kP5bg6Hw
https://www.youtube.com/embed/8p1kP5bg6Hw
Operationalization for the Machine Learning Lifecycle
Operationalization for the Machine Learning Lifecycle
The HPE ML Ops solution supports every stage of machine learning lifecycle — data preparation, model build, model training, model deployment, collaboration, ...
Jason Schroedl
Data science and ML are about more than just the code, it’s also about the data and the models - that’s one of the fundamental differences.
Jason Schroedl
The ML lifecycle is a very iterative and continuous process (e.g. retraining the models to ensure accuracy and reduce model drift or model decay) - some of the requirements are quite different than traditional DevOps.
Patrick Osborne
@jameskobielus
https://mapr.com/ebook/machine-learning-logistics/
Machine Learning Logistics | MapR
Machine Learning Logistics | MapR
How do you get a machine learning system to deliver value from big data?
NandaVijaydev
Devops concepts such as reproducibility and consistency between stages is very much the same in ML Ops. It just goes beyond deployment into monitoring and retraining and redeployment.
Jason Schroedl
But it is possible to apply CI/CD best practices and bring the speed and agility of DevOps to ML Ops. Last week HPE introduced our new ML Ops software solution focused on addressing these challenges and bringing a DevOps approach to the ML lifecycle: https://bit.ly/2lKOnBO
https://bit.ly/2lKOnBO
HPE accelerates Artificial Intelligence innovation with enterprise-grade solution for managing entire machine learning lifecycle
HPE accelerates Artificial Intelligence innovation with enterprise-grade solution for managing entire machine learning lifecycle
New HPE Machine Learning (ML) Ops solution speeds time-to-value for AI from months to days and brings DevOps agility to the ML model lifecycle
Jason Schroedl
Here’s a great interview with @NandaVijaydev about applying DevOps to machine learning: https://youtu.be/O98sNtMBaR8
https://www.youtube.com/embed/O98sNtMBaR8
Bringing DevOps to Machine Learning | HPE ML Ops
Bringing DevOps to Machine Learning | HPE ML Ops
Nanda Vijaydev, from HPE's BlueData team, discusses HPE Machine Learning Ops. HPE ML Ops brings a DevOps approach to machine learning and allows enterprises ...
jameskobielus
The MLops lifecycle is more iterative than traditional DevOps, especially in the model training loop.
Frederic Van Haren
From Big Data to Machine Learning to Deep Learning – The progress of AI http://highfens.com/2019/01/02/big-data-machine-le...
http://highfens.com/2019/01/02/big-data-machine-learning-deep-learning-artificial-intelligence/
From Big Data to Machine Learning to Deep Learning – The progress of AI - HighFens Inc.
From Big Data to Machine Learning to Deep Learning – The progress of AI - HighFens Inc.
From High-Performance Computing (HPC) to Big Data to Machine Learning to Deep Learning - this article describes the evolution and progress of AI.
jameskobielus
This is great discussion. Keep it coming. Please feel free to respond to earlier questions if you wish. Next question coming up.
Storage Godfather (HPEStorageGuy)
Here's a new YouTube video - bringing #DevOps to #MachineLearning - check it out after the chat! https://www.youtube.com/watch?v=O98sNtMBaR8
https://www.youtube.com/embed/O98sNtMBaR8
Bringing DevOps to Machine Learning | HPE ML Ops
Bringing DevOps to Machine Learning | HPE ML Ops
Nanda Vijaydev, from HPE's BlueData team, discusses HPE Machine Learning Ops. HPE ML Ops brings a DevOps approach to machine learning and allows enterprises ...
jameskobielus
Q5: What are the obstacles and challenges that prevent the operationalization of ML models in the enterprise? https://www.crowdchat.net/s/55vi8
https://www.crowdchat.net/s/55vi8

Frederic Van Haren
Compliance and Regulations should also not be underestimated. For example: GDPR
Frederic Van Haren
Cost. Speed of change.
Patrick Osborne
talking about storage and not data. We should operationalize @hpestorageguy to @hpedataguy
Jason Schroedl
Gartner estimates that 80-85% of enterprises are running into the 'last mile' problem with ML model deployment and operations. And by 2021, at least 50% of ML projects won’t be fully deployed due to lack of operationalization.
Storage Godfather (HPEStorageGuy)
@patrick_osborne Careful, don't want the branding people coming after me! I'm the Storage Godfather. Maybe I need to be the @DataGodfather?
Patrick Osborne
One challenge I have seen is proving value on too many initiatives. We have had success with BlueData on proving one workflow or service and scaling adoption incrementally
Ralph Finos
Making the results understandable in the business context so they act confidently
Abdul Matheen Raza
primarily it is lack the right tools for the operationalization and large-scale implementation of ML models. In addition, there is a lack of standards around coding, sharing, and collaboration.
Hande Sahin-Bahceci
@fvha Agreed, security and compliance teams need to get engaged early on.
Victor Ghadban
Ensuring the model was t built in a silo and that it actually solves the problem is was built for. Many business struggle to understand what was built and how to use it.
NandaVijaydev
Silos in the lifecycle is a major challenge.
NandaVijaydev
@patrick_osborne This is a complex and growing field. No longer controlled by specific vendors. Start small, prove value, and scale is applicable at every level
Hande Sahin-Bahceci
Data teams still spend too much time on operations or data preparation. #MLOPs tools and software can help #crowdchat
Dave Vellante
skills, talent acquisition...
jameskobielus
One of the chief challenges in operationalizing ML is finding enough skilled data scientists to staff 24x7 AI-driven app operations.
jameskobielus
Next question in just a minute or so.
jameskobielus
Q4: What is “operationalization”? What are the key requirements to operationalize ML models? https://www.crowdchat.net/s/35vhw
https://www.crowdchat.net/s/35vhw

Frederic Van Haren
"Operationalization" is the integration of the ML process into the daily operations of a business (products)
Patrick Osborne
deploying models at scale that demonstrate quanitifiable business value
Jason Schroedl
The dictionary defines operationalize as “to put into use” or “to make operational”. There’s a good Wikipedia entry about ML Ops at: https://en.wikipedia.org/wiki/MLOps
https://en.wikipedia.org/wiki/MLOps
MLOps - Wikipedia
MLOps - Wikipedia
MLOps (a compound of “machine learning” and “operations”) is a practice for collaboration and communication between data scientists and operations professionals to help manage production ML (or deep learning) lifecycle.[1] Similar to the DevOps or Da...
Mike Leone
Operationalization is enabling models to be deployed at scale and be seamlessly consumed by modern applications. And it's not just for better and faster insights. It's also about being able to take those insights and effectively apply them to the business.
Frederic Van Haren
Key components of operationalization are "automation" and "standardization"
Jason Schroedl
As data science teams in the enterprise move beyond initial experimentation to production deployments for ML / DL, they need to operationalize their models and the end-to-end lifecycle
Patrick Osborne
this is a great example of how we operationalized machine learning at scale, Infosight!! https://www.hpe.com/us/en/solutions/infosight.html
Abdul Matheen Raza
Operationalization in the ML context is to put the model in a DevOps like workflow with performance monitoring, versioning, etc. For operational success with machine learning, processes and tools need to support the iterative nature of the ML lifecycle
Mike Leone
Establishing effective deployment processes, accurately scoring and serving models to the right areas of the business, measuring success and failures with the right performance metrics, monitoring ongoing model performance, and identifying when to retrain a model
Jason Schroedl
It’s not just about deployment, it’s about the complete ML lifecycle. Operationalization requires a wide range of capabilities (including collaboration, reproducibility, monitoring, etc.) for the development, training, deployment, and management of operational models at scale
Victor Ghadban
models don’t serve any purpose unless they can be put to use by the business. Operationalization is the ability to take a model and integrate into some sort of workflow that can feed it data and accept responses. Most companies struggle at this point due to the complexity of it
Chris Snow
Being able to deploy ML models quickly, efficiently and in a routine way is very important - deploying and managing each ML model as "one-offs" an anti-pattern.
jameskobielus
Next CrowdChat question coming up.