$30 off During Our Annual Pro Sale. View Details »

Machine learning on Kubernetes

dharmeshkakadia
August 17, 2018
390

Machine learning on Kubernetes

Kubernetes SF meetup. Aug, 2018

Other talks at http://dharmeshkakadia.github.io/talks

dharmeshkakadia

August 17, 2018
Tweet

More Decks by dharmeshkakadia

Transcript

  1. Mobile Data Labs & Microsoft
    Dharmesh Kakadia

    View Slide

  2. Mobile Data Labs & Microsoft
    • Applied Scientist/Software Engineer,
    MobileDataLabs (acquired team inside Microsoft)
    team building AI/Analytics platform
    • Spent couple of years with Microsoft Research
    • Spent couple of years with Azure HDInsight
    • Among other things, author of “Apache Mesos
    Essentials”
    • Opinions are mine and biased
    • You can find me as @dharmeshkakadia
    everywhere
    whoami

    View Slide

  3. Mobile Data Labs & Microsoft
    Machine Learning

    View Slide

  4. Mobile Data Labs & Microsoft
    Why do you need an AI platform?

    View Slide

  5. Mobile Data Labs & Microsoft
    • Already a widely adopted standard for deploying. We are just extending it to Data/AI
    • Operational benefits for free – monitoring, CI/CD, secrets, alerts, log management ….
    • No separate process and tools for data and other parts of engineering
    • Ability to leverage latest improvements faster
    • Better resource utilization
    • Helps avoiding data silos
    • Not have to worry about installing Nvidia drivers…
    • With a caveat, that you need a little more cross functional expertise
    Why build AI platform on k8s?

    View Slide

  6. Mobile Data Labs & Microsoft
    Our AI/Analytics Platform Architecture

    View Slide

  7. Lifecycle
    Mobile Data Labs & Microsoft
    Dev
    Work inner loop
    development inside
    Juypter notebook.
    Write docker +
    YAML file when
    ready for PR
    Build
    Takes Dockerfile
    and build and
    pushes image to
    container registry
    with build tags.
    Release
    Combines YAML file
    and secrets applies
    on k8s cluster
    Monitor and
    visualize
    Produces output
    data, models and
    results. That is used
    for further
    analysis/decision
    making. No special
    ops required.

    View Slide

  8. Mobile Data Labs & Microsoft
    Example End-to-End pipeline
    • Tensorflow for model
    training and serving
    • Spark for feature
    engineering
    • Kubernetes & related tools
    for deployment
    • Data lives on blob and DW

    View Slide

  9. Mobile Data Labs & Microsoft
    • Docker as a single build tool
    • Consistent deploys – even more useful in data experiments
    • Freedom to use any library/versions I want
    • k8s as a single deployment tool
    • Easier to think about for everyone on the team
    • East to remember (and optimize!) one pattern and workflow
    • Separation of concerns
    • Build/ops tools doesn’t need to understand how TF work
    • YAML for separating code and configs.
    • Secrets for code and secrets.
    • Blobfuse for code and data paths.
    Declarative deployments

    View Slide

  10. Mobile Data Labs & Microsoft
    Volumes
    • Blobfuse
    • k8s volume plugin that makes blob data accessible as a mounted file system
    • Great for inner loop dev. Avoids additional IO to remote storage.
    • Allows read-only or read-write mounting
    • Not every tool needs to understand and integrate with blob
    • Easy when playing around with data rather than dealing with blob explorers
    • Configurable cache interval allows trading off fast access/freshness constraints.
    • Hostpath
    • Local SSD for temporary storage
    • Great for speed and intermediate results
    • Azure files
    • For permanent storage for fast non-blob data

    View Slide

  11. Mobile Data Labs & Microsoft
    Demo time !

    View Slide

  12. Serving
    • Early days.
    • Currently we store and serve directly out of blob
    • Versioned though names
    • Want to validate and understand use cases to help us to choose the right tool
    • Need something that plays nice with other data tools as well i.e. spark etc.
    • We are considering onnx and tensor-serve

    View Slide

  13. Kubeflow?
    • We evaluated very early version 0.1.0 and had a lot of issues with it
    • Opinionated and bundles a lot of tools that we currently don’t need
    • Ksonnet business is messy
    • End user simplicity is paramount
    • Start with simple tools and add tools as necessary. Don’t like big bang complex pieces
    • Too much upfront complexity, before we can get the value
    • Having said that,
    • Huge fan of the community.
    • We are keeping an eye on direction its going.
    • We do like some parts – especially TF job operator. We already use spark operator and realize the
    benefits.

    View Slide

  14. Mobile Data Labs & Microsoft
    Thanks !
    @dharmeshkakadia

    View Slide