The February playlist includes interesting episodes with Intuit, Tecton, Policygenius Inc., and an interesting chat with Duc Huba, an AI researcher and enterprise mobility solution architect consultant. Eric Dodds, our Head of Customer Success, and Kostas Pardalis, Head of Product at RudderStack, discussed the following topics:
- Migrating from on-premises to Cloud
- Demystifying AI
- MLOps and Feature stores
- Democratizing Insurance Marketplace
Check out interesting highlights from each episode below, and subscribe to get new episodes every week.
From On-Premises to Cloud with Alex Lancaster from Intuit
In this episode, Kostas and Eric interview Intuit’s Risk Data Engineering Manager, Alex Lancaster. Alex has been with Intuit for 15 years and was part of a recent massive and successful re-architecture from on-prem to cloud-based. Alex kicks off by talking about what a Data Mart is and what this looks like at Intuit. Alex says the migration from on-premises to cloud is one of the most revolutionary changes in the data engineering industry in the past 15 years because it lets you get out of the business of worrying about your hardware and your storage.
So when you move to a public cloud, and you’re in an Azure or AWS situation, those guys are investing billions every year into their cloud architecture and infrastructure. And, you know, I’m pretty sure no company, even governments, can’t compete with that kind of investment. He further dives into how Intuit moved from on-premise to cloud and how this change took them 18 months. He also shared the different cloud tools Intuit uses–Amazon EMR and Spark clusters and Parquet files, S3 and Redshift, Aurora Kinesis, Managed Kafka service, Glue, CloudWatch, and much more. He also compares security in the Cloud vs. on-premises and how Intuit has a central security team and a data handling team who have carefully architected things to a corporate standard.
One particularly interesting thing Alex shared about their successful transition from on-prem to Cloud was their decision around forklifting vs. re-stacking. Alex explained how his experience within Intuit helped him apply his software engineering skills to data engineering, and he shared how Intuit deals with data engineering challenges related to security and regulation. Eric and Kostas were shocked to know that Alex manages over 1000 pipelines in his environment!
Demystifying Artificial Intelligence with Duc Huba
In this episode, Eric and Kostas talk to Duc Haba, an AI researcher and enterprise mobility solution architect consultant. This discussion’s goal revolves around understanding artificial intelligence and why people either fear AI or place too much trust in it. He also points to how data biases that humans bring into the models can radically alter the outcome of those endeavors.
Duc kicks off with a bit about his background and his chance of working with LeVar Burton. He further continues that there is no perfect definition for AI and comes up with his version. He also highlights how people have misplaced AI fears that led to a misplaced trust in AI. Duc shares an interesting insight topic about the public and the hidden AI. He also shares how one can acquire data needed to train the AI models. Duc discussed some of the AI projects he has worked on. Finally, he anticipates AI’s near future and how he wishes to have more AI-enabled robots to do more of his boring house chores.
MLOps and Feature Stores with Willem Pienaar from Tecton
In this episode, Willem Pienaar, Tech Lead at Tecton, Eric, and Kostas discuss machine learning, features, and feature stores. Willem talks about his background from South Africa to South-east Asia, where he joined an Indonesian company, Gojek. Willem also talks about his journey from Gojek to Tecton, a company focused purely on feature stores. He shared how Tecton was built by the same team that built the Michelangelo platform at Uber.
Willem briefly explained how ‘feature’ is the data input to make predictions and explains a feature’s lifecycle. He also briefly explains the data architecture in a feature store and how feature stores evolve as a company expands. Willem shared how Tecton manages productizing complex architectures and talked about how Feast (an open-source feature store for machine learning) and Tecton work together. Finally, he talks about how Tecton is impressing VCs and preparing for a competitive, emerging market.
Democratizing the Insurance Market with Daniel Gremmell from Policygenius Inc.
Daniel Gremymell is the head of data at Policygenius, Inc., an insurance marketplace that helps people understand different options, compare quotes, and buy a policy with help from licensed experts. Daniel talks to Eric and Kostas a bit about his background and his role at Policygenius. He briefly explains what Policygenius is and how it consolidates carriers and pairs insurance customers with live experts to get the best prices and plans.
Daniel also explains how data analysts and data scientists offer improved customer experience in selecting insurance. One interesting topic Daniel shared was how roles and titles like the Head of Data are changing the industry. He says it is more around the scope and breadth of the role than the title. Lastly, he talks about organizing a company with structured embedding and also discusses Policygenius’ data stack.
Explore More Podcast Episodes and Subscribe
We release exciting new episodes every week and have a fantastic lineup of guests for the coming weeks. Listen to the latest episodes and subscribe to The Data Stack Show on iTunes or your favorite podcast player.
Powered by RudderStack
The Data Stack Show is powered by RudderStack, the smart data pipeline. With RudderStack, you can use all your customer data to answer more difficult questions and send insights to your whole customer data stack.