MLOps.community

MLOps Coffee Sessions #138 with Dattaraj Rao, Explainability in the MLOps Cycle co-hosted by Vishnu Rachakonda. // Abstract When it comes to Dattaraj's interest, you'll hear about his top 3 areas in Machine Learning. What he sees as up and coming, what he's investing his company's time into and where he invests his own time. Learn more about rule-based systems, deploying rule-based systems , and how to incorporate systems into more systems. there is no difference between ML systems and deploying models. It's just that this machine learning model is much smarter than traditional rule based models. // Bio Dattaraj Jagdish Rao is the author of the book “Keras to Kubernetes: The Journey of a Machine Learning Model to Production”. Dattaraj leads the AI Research Lab at Persistent and is responsible for driving thought leadership in AI/ML across the company. He leads a team that explores state-of-the-art algorithms in Knowledge Graphs, NLU, Responsible AI, MLOps and demonstrates applicability in Healthcare, Banking, and Industrial domains. Earlier, he worked at General Electric (GE) for 19 years building Industrial IoT solutions for Predictive Maintenance, Digital Twins, and Machine Vision. Dattaraj held several Technology Leadership roles at Global Research, GE Power, and Transportation (now part of Wabtec). He led the Innovation team out of Bangalore that incubated video track inspection from an idea into a commercial Product. Dattaraj has 11 patents in Machine Learning and Computer Vision areas. // MLOps Jobs board   https://mlops.pallet.xyz/jobs // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links Keras to Kubernetes: The Journey of a Machine Learning Model to Production book: https://www.amazon.com/Keras-Kubernetes-Journey-Learning-Production/dp/1119564832 Responsible Data Science Research | Talk @ VLDB 2022| Dattaraj Rao https://www.youtube.com/watch?v=5_19KvSiy8s Operationalizing AI/ML: Journey of an ML Model to Production | Masterclass by Dattaraj Rao https://www.youtube.com/watch?v=Zk3RiiG07Us Dattaraj Rao presenting workshop on MLOps at VISUM 2021 https://www.youtube.com/watch?v=wonUvbMDTUA Machine Learning Design Patterns book: https://www.oreilly.com/library/view/machine-learning-design/9781098115777/ --------------- ✌️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, blogs, newsletters, and more: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Vishnu on LinkedIn: https://www.linkedin.com/in/vrachakonda/ Connect with Dattaraj on LinkedIn: https://www.linkedin.com/in/dattarajrao/

MLOps Coffee Sessions #137 with Niklas Kühl, Machine Learning Operations — What is it and Why Do We Need It? co-hosted by Abi Aryan. // Abstract The final goal of all industrial machine learning (ML) projects is to develop ML products and rapidly bring them into production.    However, it is highly challenging to automate and operationalize ML products and thus many ML endeavors fail to deliver on their expectations. The paradigm of Machine Learning Operations (MLOps) addresses this issue. // Bio NIKLAS KÜHL studied Industrial Engineering & Management at the Karlsruhe Institute of Technology (KIT) (Bachelor and Master). During his studies, he gained practical experience in IT by working at Porsche in both national and international roles. Niklas has been working on machine learning (ML) and artificial intelligence (AI) in different domains since 2014. In 2017, he gained his PhD (summa cum laude) in Information Systems with a focus on applied machine learning from KIT. In 2020, he joined IBM. As of today, Niklas engages in two complementary roles: He is head of the Applied AI in Services Lab at the Karlsruhe Institute of Technology (KIT), and, furthermore, he works as a Managing Consultant for Data Science at IBM. In his academic and practical projects, he is working on conceptualizing, designing, and implementing AI in Systems with a focus on robust and fair AI as well as the effective collaboration between users and intelligent agents. Currently, he and his team are actively working on different ML & AI solutions within industrial services, sales forecasting, production lines or even creativity. Niklas is internationally collaborating with multiple institutions like the University of Texas and the MIT-IBM Watson AI Lab. // MLOps Jobs board   https://mlops.pallet.xyz/jobs // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links Website: niklas.xyz MLOps Newsletters: https://airtable.com/shrx9X19pGTWa7U3Y Machine Learning Operations (MLOps): Overview, Definition, and Architecture paper: https://arxiv.org/abs/2205.02302 --------------- ✌️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, blogs, newsletters, and more: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Abi on LinkedIn: https://www.linkedin.com/in/abiaryan/ Connect with Niklas on LinkedIn: https://www.linkedin.com/in/niklaskuehl/ Timestamps: [00:00] Niklas' preferred coffee [00:43] Introduction to Niklas Kühl [01:16] Takeaways [02:05] Subscribe to our newsletters and give us a rating here! [02:54] Niklas background [05:09] Scraping twitter data [06:58] EV's conclusions [08:24] NLP usage on Twitter [10:26] Consumer behavior production [12:03] Management and Machine Learning Systems Communication [14:00] Current hype around Machine Learning [15:10] Budgeting ML Productions [18:15] Machine Learning Operations (MLOps): Overview, Definition, and Architecture paper [22:56] Niklas' MLOps definiton   [25:55] Navigating the idea of MLOps [30:34] Return of Investment endeavor [33:58] Full stack data scientist [37:39] Defining success for different kinds of data science projects [41:06] Fun fact about Niklas [44:35] Other things Niklas do [47:02] The world is your oyster [50:57] Niklas' day to day  life [52:48] One lecture Niklas can drop in [53:57] Foundational models [58:20] Wrap up

MLOps Coffee Sessions #136 with Andrew Dye, Systems Engineer Navigating the World of ML co-hosted by David Aponte. // Abstract We don't hear that much about working at a very low level on this podcast but they are still very valid. Andrew is able to give us his take on why and what you need to keep in mind when you are working at these low levels and why it is very important when you are a Machine Learning Engineer and how the two can play together nicely. Most MLOps teams are formed using existing people and exitsing engineers. More often than not you have to blend these various disciplines and it works well when there's a common goal. // Bio Andrew is a software engineer at Union and contributor to Flyte, a production grade data and ML orchestration platform. Prior to that he was a tech lead for ML Infrastructure at Meta, where he focused on ML training reliability. // MLOps Jobs board   https://mlops.pallet.xyz/jobs // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links --------------- ✌️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, blogs, newsletters, and more: 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/ Connect with Andrew on LinkedIn: https://www.linkedin.com/in/andrewwdye Timestamps: [00:00] Andrew's preferred coffee [03:30] Introduction to Andrew Dye [03:33] Takeaways [07:32] Huge shoutout to our sponsors UnionML and UnionAI! [07:48] Andrew's background [10:08] Andrew's learning curve [11:10] Bridging the gap between firmware space and MLOps [12:18] In connection with Pytorch team [12:54] Things that should have learned sooner [14:54] Type of scale Andrew works on [17:42] Distributed training at Meta [19:55] Managing the huge search space [22:18] Execution patterns programs [23:20] Non-ML engineers dealing with ML engineers having the same skill set [26:44] Pace rapid change adoptation [29:18] Consensus challenges [32:26] Abstractions making sense now [34:53] Comparing to others [39:21] General principles in UnionAI tooling [41:54] Seeing the future [43:54] Inter-task checkpointing [44:52] Combining functionality with use cases [46:17] Wrap up

MLOps Coffee Sessions #135 with Sasha Ovsankin and Rupesh Gupta, Real-time Machine Learning: Features and Inference co-hosted by Skylar Payne.   // Abstract Moving from batch/offline Machine Learning to more interactive "near" real-time requires knowledge, team, planning, and effort. We discuss what it means to do real-time inference and near-real-time features when to do this move, what tools to use, and what steps to take.   // Bio Sasha Ovsankin Sasha is currently a Tech Lead of Machine Learning Model Serving infrastructure at LinkedIn, worked also on Feathr Feature Store, Real-Time Feature pipelines, designed metric platforms at LinkedIn and Uber, and was co-founder in two startups. Sasha is passionate about AI, Software Craftsmanship, improvisational music, and many more things.   Rupesh Gupta Rupesh is a Sr. Staff Engineer in the AI team at LinkedIn. He has 10 years of experience in search and recommender systems.   // MLOps Jobs board   https://mlops.pallet.xyz/jobs // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links   --------------- ✌️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, blogs, newsletters, and more: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Skylar on LinkedIn: https://www.linkedin.com/in/skylar-payne-766a1988/ Connect with Sasha on LinkedIn: https://www.linkedin.com/in/sashao/ Connect with Rupesh on LinkedIn: https://www.linkedin.com/in/guptarupesh Timestamps: [00:00] Sasha's and Rupesh's preferred coffee [01:30] Takeaways [07:23] Changes in LinkedIn [09:21] "Real-time" Machine Learning in LibnkedIn [13:08] Value of Feedback [14:24] Technical details behind getting the most recent information integrated into the models [16:53] Embedding Vector Search action occurrence [18:33] Meaning of "Real-time" Features and Inference [20:23] Are "Real-time" Features always worth that effort and always helpful? [23:22] Importance of model application [25:26] Challenges in "Real-time" Features [30:40] System design review on Pinterest [36:13] Successes of real-time features [38:31] Learnings to share [45:52] Branching for Machine Learning [48:44] Not so talked about discussion of "Real-time" [51:09] Wrap up

MLOps Coffee Sessions #134 with Jeremy Thomas Jordan, Building Threat Detection Systems: An MLE's Perspective co-hosted by Vishnu Rachakonda. // Abstract There is a clear pattern that we have been seeing with some of these greats in MLOps. So many use writing as a forcing function to learn about where they have holes in their understanding of something.    If you are not writing, this episode is important as to why writing is important for your own development. Jeremy goes into writing in depth as to how beneficial it is for him to write and for him to see that he doesn't understand something if he cannot re-articulate it in writing. // Bio Jeremy is a machine learning engineer currently working at Duo Security where he focuses on building ML infrastructure to operate threat detection systems at scale. He previously worked at Proofpoint, where he built models for phishing and malware detection. // MLOps Jobs board   https://mlops.pallet.xyz/jobs // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links Website: https://www.jeremyjordan.me/ --------------- ✌️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, blogs, newsletters, and more: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Visnu on LinkedIn: https://www.linkedin.com/in/vrachakonda/ Connect with Jeremy on Twitter: https://twitter.com/jeremyjordan

MLOps Coffee Sessions #133 {Podcast BTS} with Chip Huyen, Real-time Machine Learning with Chip Huyen co-hosted by Vishnu Rachakonda. // Abstract Forcing functions and how you can supercharge your learning by putting yourself into a situation where you know you either have a responsibility to others to learn or accountability on you so you have to learn.   It's not that hard when you think about streaming machine learning. It's not that big of a mental barrier to cross. It is simple in theory but maybe it's more complicated in practice and that's exactly where Chip's perspective is. // Bio Chip Huyen is a co-founder of Claypot AI, a platform for real-time machine learning. Previously, she was with Snorkel AI and NVIDIA. She teaches CS 329S: Machine Learning Systems Design at Stanford. She’s the author of the book Designing Machine Learning Systems, an Amazon bestseller in AI. // MLOps Jobs board   https://mlops.pallet.xyz/jobs // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links Landing page: https://claypot.ai Designing Machine Learning Systems book: https://www.amazon.com/Designing-Machine-Learning-Systems-Production-Ready/dp/1098107969 --------------- ✌️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, blogs, newsletters, and more: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Vishnu on LinkedIn: https://www.linkedin.com/in/vrachakonda/ Connect with Chip on LinkedIn: https://www.linkedin.com/in/chiphuyen/

MLOps Coffee Sessions #132 {Podcast BTS} with Ian Schweer, What is Data / ML Like on League? co-hosted by Skylar Payne.   // Abstract If you're not an avid gamer yourself, you have never really seen how ML might be used in the gaming space. It's so interesting to see the things that are different like full stories of players' games from start to finish.   // Bio On the surface, Ian is an excellent developer who gets things done. Underneath, he is much more. Ian is a reliable and trustworthy teammate who demonstrates an exceptional ownership mentality.   Here's a fair share of Ian's job history: 2014 - UCI (With Skylar!) 2015 - Adobe Primetime (SWE) 2017 - Adobe Product and Customer Analytics (SWE) 2019 - Doordash Data Infra (SWE) Current - Riot Games on League   // MLOps Jobs board   https://mlops.pallet.xyz/jobs // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links Landing page: https://www.riotgames.com/en --------------- ✌️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, blogs, newsletters, and more: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Skylar on LinkedIn: https://www.linkedin.com/in/skylar-payne-766a1988/ Connect with Ian on LinkedIn: https://www.linkedin.com/in/ianschweer/ Timestamps: [00:00] Ian's preferred coffee [02:10] Takeaways [05:14] Please hit the like button and leave us a review. Please subscribe also! [05:45] Engineering Community Mental Health Awareness [07:33] Coping mechanism [09:29] Increase in video game playing   [11:20] Ian's career progression [17:55] Lessons to apply in the Data space [24:23] Challenges at Riot [34:18] Real-time element [39:09] Ian's day-to-day responsibilities [43:13] Analysis vs. Production Code Quality [48:11] Tools and techniques on the reality of writing production codes [55:00] What would you change your career into? [57:00] Ian's best practices advise [58:28] Ian's favorite video game [59:58] Wrap-up

MLOps Coffee Sessions #131 {Podcast BTS} with Ethan Rosenthal, Let's Continue Bundling into the Database co-hosted by Mike Del Balso. // Abstract The relationship between ML Engineers and Product Managers is something that we don't talk about enough. We've got to get this right. If we don't get this right, either you're not focusing on the business problems in the right way or the Product Managers are not going to understand the tech appropriately to help make the right decisions. // Bio Ethan works on the Conversations Team at Square leading a team of Artificial Intelligence Engineers. Ethan's team builds applied AI solutions for Square Messages, a messaging hub for Square merchants to communicate with their customers. Prior to Square, Ethan spent time as a freelance data science consultant building machine learning products for a range of companies, from pre-seed startups to Fortune 100 enterprises.    Ethan got his start in data science working at two different e-commerce startups, Birchbox and Dia&Co. Before data science, Ethan was an actual scientist and got his Ph.D. in experimental physics from Columbia University. // MLOps Jobs board   https://mlops.pallet.xyz/jobs // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links https://www.ethanrosenthal.com/ Relevant blog posts:   https://www.ethanrosenthal.com/2022/05/10/database-bundling/ https://www.ethanrosenthal.com/2022/07/19/materialize-ml-monitoring/ https://www.ethanrosenthal.com/2022/01/18/autoretraining-is-easy/ --------------- ✌️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, blogs, newsletters, and more: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Mike on LinkedIn: https://www.linkedin.com/in/michaeldelbalso/ Connect with Ethan on LinkedIn: https://www.linkedin.com/in/ethanrosenthal/ Timestamps: [00:00] Ethan's preferred coffee [00:10] Introduction to co-host Mike Del Balso [00:43] Takeaways [04:10] Sign up for our newsletter! [04:47] Ethan's team [06:49] Ethan's team improvement [08:40] Product manager role at Square [10:39] Large Language Models [12:22] Big questions to figure out [15:45] Cost of false-positive [18:20] Company Vocabulary [20:11] MLOps concerns and challenges around Large Language Models [23:42] Data learning management [27:36] Leveling trade-offs [30:25] Ethan's Database Bundling blog [34:32] Feature Stores [38:24] Streaming databases [41:57] Best of both worlds trade-off highlight [43:51] Rosenthal data [46:40] Ethan's freelancing [47:46] Risk-reward trade-off [49:17] Ethan as a professor [51:14] Wrap up

MLOps Coffee Sessions #130 {Podcast BTS} with Andrew Yates, Adversarial MLOps on Other People's Money: Lessons Learned from Ad Tech co-hosted by Abi Aryan. // Abstract Design ML to be components in a larger system with stable interfaces is not tracible to monitor the entire stack as a black box. You need intermediate ground-truth signals. Ads are designed in this way. You can go from profitable to non-profitable real quick with ads. This will determine whether your company is around a year or two. You play with money and sometimes you play a lot of it so make sure that it's correct. // Bio Andrew Yates formerly led ads ranking, auction, and marketplace engineering and research teams at Facebook and Pinterest. He specializes in designing billion-dollar content marketplaces that maximize long-term revenue while protecting both seller and user experiences. Andrew has published over a dozen patents in online advertising optimization. // MLOps Jobs board   https://mlops.pallet.xyz/jobs // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links --------------- ✌️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, blogs, newsletters, and more: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Abi on LinkedIn: https://www.linkedin.com/in/abiaryan/ Connect with Andrew on LinkedIn: https://www.linkedin.com/in/andrew-yates-0217a985/

MLOps Coffee Sessions #129 {Podcast BTS} with Catherin Breslin, Voice and Language Tech co-hosted by Adam Sroka. // Abstract Back in the day, Speech Recognition was its own thing. It's a very different flavor of Data Science. You could not use a lot of the tools. It wouldn't cross over to this type of machine learning. Now, with the advancements, Speech Recognition and Machine learning are coming in together. It's interesting to hear right from someone with a Ph.D. level working with some of the biggest companies in the world doing it. The fact that something like Alexa is lots of models back to back and just fathom the complexity of that is quite cool! // Bio Catherine is a machine learning scientist and consultant based in Cambridge UK, and the founder of Kingfisher Labs consulting. Since completing her Ph.D. at the University of Cambridge in 2008, Catherine has commercial and academic experience in automatic speech recognition, natural language understanding, and human-computer dialogue systems, having previously worked at Cambridge University, Toshiba Research, Amazon Alexa, and Cobalt Speech. Catherine has been excited by the application of research to real-world problems involving speech and language at scale. // MLOps Jobs board   https://mlops.pallet.xyz/jobs // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links www.catherinebreslin.co.uk https://catherinebreslin.medium.com/ MLOps Community Newsletter: https://airtable.com/shrx9X19pGTWa7U3YTwitter: https://twitter.com/catherinebuk --------------- ✌️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, blogs, newsletters, and more: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Adam on LinkedIn: https://www.linkedin.com/in/aesroka Connect with Catherine on LinkedIn: https://www.linkedin.com/in/catherine-breslin-0592423a/ Timestamps: [00:00] Catherine's preferred coffee [01:50] Takeaways [03:59] Introduction to Catherine Breslin [05:04] Subscribe to our newsletter! [06:25] Catherine's background [08:13] Speech Recognition trajectory [09:36] Challenges around technologies and tools [11:34] Reflective trend [13:02] Developer experiences hiccups [15:09] Speech Recognition use case backup [16:56] Toshiba research [17:48] Transition from a research lab to working in the industry [20:01] Unit test of Speech Recognition [20:56] Alexa [22:33] Maturity process of Speech Recognition [26:48] Speech Recognition unrecognizing challenges [30:38] Mechanical Terk [33:00] Social media listening [34:05] Pipeline models and speed of Speech Recognition [36:48] Development of Speech Recognition excited about [37:23] Data from people for the Speech Recognition system vs Scowering news vs watching Youtube for a long time [40:00] Disappearing Languages [41:30] Future of an online practice partner [43:17] Speech-to-speech translation [44:04] Interesting ways to use unfamiliar models to achieve a result [45:40] Meeting transcriptions [48:37] First toy problem of a new Speech Recognition learner [51:37] Kingfisher Labs' problems to tackle [52:18] Off-the-shelf solution [53:38] Translation layer [54:15] Connect with Catherine on Twitter and LinkedIn for available jobs [54:43] Wrap up