Your ML Engineering Needs

We asked what you wanted to hear next on our Coffee sessions and the vote was in favor of feature stores! Today the usual suspects Demetrios Brinkmann and David Aponte sat down to talk with Jim Dowling CEO of Logical Clocks and Venkata Pingali CEO of scribble data to talk about feature stores, what they are, why we need them, some business implications and everything in between! As always if you enjoyed the session let us know or reach out to us on slack! Check out what Jim is doing around Hopsworks and open-sourced feature stores at Logical Clocks: https://www.logicalclocks.com/ Find out more about the feature stores that Venkata is building at Scribble Data: https://www.scribbledata.io/ --------------- ✌️Connect With Us ✌️ ------------- Join our slack community: https://go.mlops.community/slack Follow us on Twitter: @mlopscommunity Sign up for the next meetup: https://go.mlops.community/register Catch all episodes, Feature Store, Machine Learning Monitoring and Blogs: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with David on LinkedIn: https://www.linkedin.com/in/aponteanalytics/

MLOps Community Meetup #6 Mid-Scale Production Feature Engineering In the 4th online meetup for our MLOps.community We spoke with Dr. Venkata Pingali, CEO of Scribble Data and Feature engineering expert! Scribble Data helps build and operate production feature engineering platforms for sub-fortune 1000 firms. The output of the platforms is consumed by data science and analytical teams. In this talk, we discuss how we understand the problem space, and the architecture of the platform that we built for preparing trusted model-ready datasets that are reproducible, auditable, and quality checked, and the lessons learned in the process. We will touch upon topics like classes of consumers, disciplined data transformation code, metadata and lineage, state management, and namespaces. This system and discussion complement work done on data science platforms such as Domino and Dotscience. Dr. Venkata Pingali is Co-Founder and CEO of Scribble Data, an ML Engineering company with offices in India and Canada. Scribble’s flagship enterprise product, Enrich, enables organizations to address 10x analytics/data science use cases through trusted production datasets. Before starting Scribble Data, Dr. Pingali was VP of Analytics at a data consulting firm and CEO of an energy analytics firm. He has a BTech from IIT Mumbai and a Ph.D. from USC in Computer Science. This was a virtual fireside chat between Dr. Venkata Pingali, Demetrios Brinkmann, and the MLOps community. Relevant links can be found below. Join our MLOps slack community: https://bit.ly/3aOTwgR Catch all episodes, Feature Store, Machine Learning Monitoring and Blogs: https://mlops.community/ --------------- ✌️Connect With Us ✌️ ------------- Join our slack community: https://go.mlops.community/slack Follow us on Twitter: @mlopscommunity Sign up for the next meetup: https://go.mlops.community/register Catch all episodes, Feature Store, Machine Learning Monitoring and Blogs: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Dr. Venkata Pingali on Linkedin: https://www.linkedin.com/in/pingali/

Anecdotally, only 2% of the models developed are productionized, i.e., used day to day to improve business outcomes. Part of the reason is the high cost and complexity of productionization of models. It is estimated to be anywhere from 40 to 80% of the overall work. In this talk, we will share Scribble Data's insights into productionization of ML, and how to reduce the cost and complexity in organizations. It is based on the last two years of work at Scribble developing and deploying production ML Feature Engineering Platform, and study of platforms from major organizations such as Uber. This talk expands on a previous talk given in January. First, we discuss the complexity of production ML systems, and where time and effort goes. Second, we give an overview of feature engineering, which is an expensive ML task, and the associated challenges Third, we suggest an architecture for Production Feature Engineering platform. Last, we discuss how one could go about building one for your organization More details: https://confengine.com/odsc-india-2019/proposal/9675 Conference Link: https://india.odsc.com

This talk addresses the following questions: 1. What should a production ML feature engineering platform have and why? 2. When do I need one? 3. What are my options if I have to build one? This talk draws upon the Scribble’s experience in building and evolving a production feature engineering platform, and the many conversations we have had with user data scientists. The talk will focus on the learnings, and not on the Scribble product itself, and expand on the talk from Fifth Elephant Mumbai in Jan 2019 on reducing costs. Dr. Venkata Pingali is Co-Founder and CEO of Scribble Data, an ML Engineering company based in Bangalore and Denver. Scribble’s flagship enterprise product, Enrich, accelerates ML productionization in enterprises. Before starting Scribble Data, Dr. Pingali was VP of Analytics at a political data consulting firm. He has a BTech from IIT Mumbai and a PhD from USC in Computer Science.

ML models have to be both economically viable and FAccT (Fair, Accountable, Transparent). The terminology is new but not the need to defend models or to attest they can be trusted. Such requirements were present from the 70s for credit scoring models. What has changed is the scale and scope. This session is about: 1. Current status of the FAccT conversation and methods (legal, consumer expectations etc.) 2. What are the compelling reasons for an organisation to invest in FAccT? What it takes to achieve end-to-end FAccT? 3. Practical challenges in implementing FAccT both at organisational and at technical level 4. Various efforts and initiatives around the world (academic, industry-related, products/tooling) Session details here: https://hasgeek.com/fifthelephant/operationalizing-responsible-ml/

The ‘science’ in Data Science refers to the process of developing systematic understanding of the world through observations and experimentation. This science is happening in the context of fast moving organizations, in near realtime, and by folks who have varied backgrounds. The most familiar version of the experimentation is the A/B testing. In this session, panelists Paul Meinshausen, Co-Founder at Aampe, and Bargava Subramanian, Co-Founder and CTO of Binaize explain how organizations can think about experimentation, and factors such as feasibility, value and ROI. For more details, visit: https://hasgeek.com/fifthelephant/making-data-science-work-session-5/

In this session, Dmitry Pretrov and Ivan Shcheklein - co-founders of DVC - discuss: 1. How software engineering principles apply to Machine Learning development and deployment. 2. How ML systems are different from traditional applications. 3. Importance of data versioning. For more details, visit: https://hasgeek.com/fifthelephant/making-data-science-work-session-3/