درباره این اپیزود
Real estate data is a classic example of “wide data” at scale. In this episode, Paras Doshi, Sr Director of Data & Insights at Opendoor, explains how his team rebuilt their data platform to support pricing homes using 500+ features per transaction across MLS data, third-party sources, and millions of proprietary photos.
The conversation dives into Opendoor’s four intelligence layers for home valuation, the move from fragmented team-owned stacks to a centralized platform, and how deep lineage, monitoring, and data quality became critical as AI turned the default strategy.
The episode also explores what enabled this transformation beyond technology alone: investing in strong data foundations and having leadership that meets teams in the middle, creating the conditions to deliver real business value with AI.
Chapters:
00:00 - Introduction
06:02 - Opendoor’s four layers of intelligence for home valuation
08:59 - Wide data vs. big data: 500+ data points to power home valuation
11:30 - Infrastructure upgrade: consolidating fragmented stacks
26:10 - Default-to-AI strategy using a POC-first approach
34:21 - Top-quartile hiring: talent strategy for 100x engineers
Paras Doshi leads Decision Science and Data Engineering at OpenDoor. His background spans hands-on and leadership data science, analysis, and engineering roles across industries. Paras is a Fellow of the Institute of Analytics, a data science and AI researcher, and an active blogger with 600+ posts and over 2M reads.
Roy Daniel is the Co-Founder & CEO of definity, the optimization and observability platform for the Lakehouse and Spark ecosystem.
Subscribe to Data Innovators for more conversations with technical data leaders solving complex challenges across the data ecosystem.
#DataInnovators #DataEngineering #DataPlatforms #PropTech#AIStrategy #Opendoor #Podcast #definity
