techblog.realtor.comrealtor.com Tech Blog - Technology and Product blog at

Realtor.comVisit realtor.com realtor.com Tech Blog Technology and Product blog at Realtor.com Menu Home Engineering Careers Home Made How Realtor.com Adopted In-house Feature Drift Tooling Built on Open Source What’s one of the biggest problems that always seems to be an afterthought when it comes to building then releasing machine-learning (ML) models? From my perspective, it’s ensuring validated” models continue to work as they should for an indeterminate amount of time after being deployed into production. ML teams spend months validating data queries, building and evaluating modeling options, determining the optimal model to use in production, and building dashboards to monitor key metrics. Unfortunately, this same effort is usually not given to data and model monitoring. Shifts in data and model prediction distributions can lead to model performance degradation. As economic and societal factors change consumer behavior, the new data fed into ML models can move further away from the training data. This is even more pronounced in the cyclical real estate industry with many external factors, such as interest rates, that heavily impact consumer behavior. In this post, I’ll present an overview of an in-house data and model monitoring solution using the open-source package, Evidently . We will also walk through the necessary code to set up a feature drift pipeline using Evidently and another open-source library, Metaflow . Continue reading How Realtor.com Adopted In-house Feature Drift Tooling Built on Open Source” Author Evan Lutins Posted on April 15, 2024 April 16, 2024 Categories Engineering Tags 2024 , Feature Monitoring , machine learning , Model Deployment , Model Draft , Open Source , Production ML , Realtor.com , Technology Augmented Gallery with Property Details Makes Searching for the Perfect Home Easier Photos are critical to the home-buying experience. Imagine doing a home search without them! How would you know if a home met your expectations if you couldn’t actually see what it looked like? At the same time, descriptions contextualize what you see in a photo. It’s helpful to know whether a small room with shelves is a pantry or a walk-in closet or if the ceilings are as high as they appear in the photos. Is that gray countertop granite or quartz? Is the bathroom tile from Home Depot or Italy? These details matter. In the interest of making it as easy as possible for home buyers to find the homes of their dreams, the For Sale Listing Detail Page (LDP) team at Realtor.com decided to marry photos with text in homebuyers’ search results to provide an augmented gallery experience for browsing and evaluating available homes. Unique to the industry, this new gallery feature provides a delightful, integrated consumer experience that simplifies and enriches the process of finding a home. Property data provided by the multiple listing services (MLS) is presented in the image gallery, allowing users to see relevant information and understand images in their proper context. This feature represents the largest surveyed opportunity for increasing preference from our current LDP, specifically for non-Realtor.com users and first-time users. By introducing contextually relevant data to the gallery experience, we can assist users in evaluating whether a home fits their needs. Continue reading Augmented Gallery with Property Details Makes Searching for the Perfect Home Easier” Author Tonja Hardyman Posted on December 20, 2023 December 20, 2023 Categories Product Tags 2023 , AI , machine learning , Product , Realtor.com , Technology Using Helm Library Charts to Improve Developer Experience With Tyk Operator Let’s start off with a hot take. Documentation is not as important as you think it is. Before you put this article down and write me off, hear me out…it gets better. I am not saying documentation isn’t important, just that it might not be the most efficient way to improve Developer Experience. As programmers, we don’t relish the thought of writing documentation, but we love to tell people to RTFM. What if I told you there was a way to write less documentation but also have fewer user support questions coming your way? Sounds pretty good, right? So, what is it? The answer is good design. A good design can be worth a thousand words…in documentation. I love the way that this is summed up in the following meme from Fernando Villalba: Continue reading Using Helm Library Charts to Improve Developer Experience With Tyk Operator” Author Cabe Waldrop Posted on December 15, 2023 December 15, 2023 Categories Engineering Tags 2023 , API , Design , Helm , Kubernetes , Platform Engineering , Realtor.com , Technology , Tyk From Flutter to Full Native: How We Optimized Performance and User Experience for Realtor.com’s Mobile App You may be familiar with Flutter, an open-source UI software development kit from Google that’s useful for developing cross-platform mobile apps from a single codebase for any web browser. It was released in May 2017, and we use it to develop many of the screens on our mobile app. While this solution worked well back then, today’s users expect more—more accessibility, more performance, more features, more animation, and more innovation. To that end, about two years ago, we decided to invest heavily in a mobile-first approach to increase the growth and adoption of our mobile app. As we started the research and defined the problem to solve, we asked ourselves: What is the right technology that will help us achieve our vision of being mobile-first while providing a high-performance, native experience? How do we ensure the solution can evolve as the OS ecosystem changes and differentiate our offering from other available products? Continue reading From Flutter to Full Native: How We Optimized Performance and User Experience for Realtor.com’s Mobile App” Author Realtor.com Mobile team Posted on August 29, 2023 August 29, 2023 Categories Engineering Tags Flutter , Mobile , Native , Realtor.com , Technology Improving Deep Link Stability in the Realtor.com Android App Searching for a home is often not a solo endeavor. Many people, including family, friends, and real estate agents, can be involved. Keeping everyone informed involves a lot of communication, and it’s common to share links to listings of homes and properties with others who are on the journey with you. At Realto r.com, we use a third-party SDK called AppsFlyer to generate deep links—hyperlinks that lead to specific content within our app, such as a home listing or saved search. AppsFlyer provides a convenient way to create short deep links that can be easily shared via messaging apps, email, and other communication channels. If a recipient doesn’t have Realtor.com installed, the link will take them to the particular listing or search page on the Realtor.com website, which further has a banner at the top to lead the user to the Google Play store so they can quickly download it to view the linked content. It’s usually a seamless process. However, when deep links aren’t working, it can be highly frustrating for users. Continue reading Improving Deep Link Stability in the Realtor.com Android App” Author Puneet Chugh Posted on July 25, 2023 July 24, 2023 Categories Engineering Tags 2023 , Android , Application , AppsFlyer , DeepLink , Mobile , Realtor.com , Technology A Day in the Life of a Design Director at Realtor.com: Marisa Parsons, Senior Director of Product Design For Marisa Parsons , developing her team is at the heart of her work at Realtor.com. I’m passionate about helping our designers grow in their careers, and I’m excited to coach them and create opportunities for them to further their skills,” she says. It’s our job to solve complex problems and deliver work with the highest level of craft. The best way for us to do that is to keep learning and growing.” Continue reading A Day in the Life of a Design Director at Realtor.com: Marisa Parsons, Senior Director of Product...