How to CTO: Dan Mathews of ROOM8 Shares how the Co-Living Data & AI Platform was Built


Dan Mathews
CTO, Room8



How does a young technology company succeed in a crowded market? It’s easier said than done, but the best way to achieve that goal is to disrupt the industry by providing innovative solutions to your customers. That’s why ROOM8’s Chief Technology Officer Dan Mathews believes the job description of a CTO requires him to achieve two goals.

  1. Increase the financial value of any technical asset, meaning the public and private uses of hardware and software.
  2. Refer to the former and add, “at all times.”

 

Dan says, however, “the job is different than the profession.” 

The *profession* requires practicing the first goal of a CTO’s job, but in a way that’s consistent with the company’s mission and overall ethics in all communities. The job and the professional standards are always on top-of-mind for Dan as he works throughout his day.

Beyond the role and profession of a CTO, Dan says all tech companies, new and seasoned, should create services or products based on the answers of these questions: 

  1. Is your product or service addressing the urgent needs of your target market?
  2. If so, did your technology or product improve their quality of life?

 

If the answer is yes for both, Dan says, that’s the difference between disruptive, transformative innovation and familiar and faddish revivals of tried and true methods.

THE PRODUCT

ROOM8 is more than a data and AI company, says Dan—it’s an IoT company, a community, a solution to a prevalent solution to a problem young people face today: a one-stop-shop for all things related to relocation and co-living.

Once users download the app, they select their desired location, move-in date, max-min budget, and their interests—art, food, outdoors, parties, movies, sports, books, social settings. Next, they answer some yes or no questions:

  1. Clean freak?
  2. Night owl?
  3. Smoking?
  4. Pets?
  5. Furniture rental through CORT?

 

Then, users input their date of birth, schooling, occupation, gender, profile picture and a 40-character bio asking to “describe yourself.” And boom. Users are left with a home screen full of roommate candidates to choose from that have similar interests, budget, move-in dates and location preferences—all without having to share your personal information. If users find a roommate who seems to be a match, instead of having to meet in-person or exchange phone numbers to chat, they can chat or video-chat within the app itself. 

If all goes well, the new roommates don’t even have to leave the app to look for places they’d like to rent. ROOM8 partnered with one of the nation’s largest apartment owners and operators to gain access to their Enterprise Resource Planning (ERP) property management systems. That means tens of thousands of apartment inventory in all major metropolitan cities in about 23 states are available in real-time with accurate pricing and availability dates on the app itself. There is also a function in ROOM8 where independent landlords can post their openings on the platform, which CEO Jessica Chen told QWORK takes up about 5% of ROOM8’s inventory.

THE BACK STORY

Dan Mathews told QWORK he likes to describe his education as “problem-solving.” In more literal terms, Dan studied mathematics and philosophy at the University of Arizona as an undergrad with a short stint in graduate school before discontinuing. You might be wondering how the CTO of a disruptive tech start-up would describe his collegiate studies as “problem-solving,” if philosophy was a subject he spent time and money on to study. Dan thinks, however, STEM and philosophy-ethics studies go hand-in-hand if you’re trying to innovate.

Problem-solving, Dan said, is “not just about the language that you formulate to solve the problem—it’s also about all the languages that you could possibly think about this… So even if you solve a problem, that problem in that particular language is only relevant and Germain to the speakers of that language.”

With the development of software and technology only exponentially increasing from its initial boom in the mid-to-late 2000s, he consciously and intuitively made the choice to be a software engineer with a specialty in math. This combination allowed him to seamlessly start to understand the data science aspect of software engineering—a skill Dan deems as “critical,” to CTO-ing. Throughout his career, he had leadership opportunities at different companies, including in the technology wing of CSAA, where the first iteration of ROOM8 originated. There, he focused on developing the proprietary matching algorithms that make ROOM8 the AI and data platform it is. 

When ROOM8 became an independent start-up, Dan came on as a co-Founder and CTO and the company took CSAA’s idea of ROOM8 as solely a roommate-matching app and later added the apartment-inventory-search capabilities. “The next step is now, how are we powering that co-living experience in real-time?” Dan also added that Covid had shined “a larger spotlight on this gap in software and co-living.”

The coronavirus pandemic illustrated, yet again, how access to software determines a certain quality of life. We live in the narrow prism of a pandemic where it’s hard to physically interact or visit possible apartments and ROOM8’s made it possible for users to safely, “get everything that they need in a secure way that maintains their privacy and confidentiality.”

THE NITTY-GRITTY

Dan and his team have set-up a data-and-marketplace infrastructure to securely and reliably powers the renter’s journey. But what gives them a competitive advantage? They didn’t develop ROOM8 to address only one aspect of co-living—they’re developing ROOM8 to address solutions to co-living holistically. Here’s how…

Dan shared it took about one quarter to get ROOM8 up and running back in October of 2019. With a platform as large as ROOM8’s, the CTO broke up the work into three technical regions:

  1. Platform
  2. Data
  3. Product

 

Having this three-tier strategy allows Dan and his team to separate and focus on how the three interact appropriately from technical and commercial objectives.

The platform piece, Dan defined, is the cloud and on-premises environments used to drive the transmissions received from their partners. The CTO shared they engineered their platform, “to bring information from the Internet of Properties as useful data to the Internet of Roommates,” which together makes up the co-living IoT.

To understand how this platform was optimized, it’s important to learn the reason *why* it’s important to ROOM8 to set it up the way the tech team did. “This hybrid IoT/AI approach will hopefully create a stronger connection between the Proptech—Internet of Properties—and the Fintech—Internet of Roommates—ecosystems to raise the bar for the co-living experiences of the future.” ROOM8’s mission, as Dan also shared, is for every roommate to have or have access to ROOM8 and with that option, users are able to achieve everything they need to do in co-living. 

In order to achieve that mission, it behooves ROOM8 to marry the two silos—the Internet of Properties and the Internet of Roommates—together as *the* co-living technology.

As far as *how* Dan and his team work towards optimizing the platform, one general method he shared is to approach it from an AI perspective, “to design coliving algorithms that use historical information to improve [the] performance of a coliving task.” Dan added, “so, co-living tech amounts to improving unsupervised learning tasks between distributed IoT systems.”

An entire quarter is dedicated to optimizing the data piece with the Internet of Things (IoT) and their AI models. Branches under AI Dan and his team work with on a daily basis include machine learning and under machine learning is deep learning.

An example of these branches of computer and data science at work is this technological concept ROOM8’s trying to implement called ambient computing. To put in laymen’s terms, ambient computing for ROOM8 means accessing the benefits and solutions without the reliance on a desktop or an app on a smartphone. 

When you consider ambient computing, you have to also consider IoT as a whole and how ROOM8’s IoT—a combination of partners’ information, customers’ information and how ROOM8 packages those pieces of information to achieve their services and products—connects and interacts with each other. This data layer, Dan said, is “very sensitive and very time-dependent.” It depends on how people want to work with certain technologies, and what information they’re willing to share or keep private. 

Data privacy and security are of the utmost importance to the ROOM8 tech team and ROOM8 as a whole. They recognize, “all information is valuable,” and believe if anything happens with your information, “it must be articulated to you in a simple way.” An example of how this philosophy is put into practice is when users sign up on the app, a message that pops up before you proceed that simply say, “we’ll never sell or share your information.” Instead of scrolling through a long privacy policy and blindly or cautiously clicking “accept,” you’re given reassurance with a sentence your data is safe. 

The duration of a project like this, Dan said, is not something you can plan a start and end date for because “that data layer is constantly evolving.” There might be another solution that arises within the industry that serves the customers a lot better over time, so there’s always an incentive to build upon and improve the foundation already established.

The final piece Dan and his team are focused on is the product—Dan describes the product as the ROOM8 apps, services and SAS that people interact with on a daily basis. The coding languages used most in building ROOM8 are Java, Python and Java Script, as well as Rust and Wolfram Mathematica for “numerical”, “symbolic” and “rule-based” things. For example, Dan explained, an AI recommendation model—like where you type in some things you like and you get a recommendation based on your inputs— can be developed with Mathematica because the tools around Mathematica and Mathematica itself has a lot of the natural language processing and that natural language understanding that goes into that input. 

You can bet that Dan is focused on all three layers—platform, data and product—every single day with each day starting at around 6:00 AM PST.

THE DAY-TO-DAY

ROOM8 has had its employees work from home since the start of Covid. These days, he’s either working from home or remotely working from a coffee shop with his mask on and staying socially distant from others. The day’s work is made up into four quarters:

    1. Road-mapping and coaching
    2. Developing and working on his own projects
    3. Analyzing his work, clients’ expectations and the work of others on his team
    4. Exploring new topics and potentially implementing new solutions

The first quarter of the day for Dan as CTO is checking his destinations and prioritizing *how* and *where* he can direct and coach his team to work. That first task takes about 5% of his day where he’s gauging the team’s progress is overall on their Objectives and Key Results (OKRs). The, where, when, how and industry standards/feedback and customer feedback are aspects thought about in this stage. 

This road-mapping-and-productivity check-in replaces daily, early morning stand-up team meetings with weekly stand-up meetings instead. Dan believes it’s the CTO’s responsibility to streamline the technology and tools and have his talented engineers and data scientists focus on “the management of best implementation.” Therefore, they’re focusing on doing their best work.

The next 20% of his day is dedicated to the “coaching” aspect of CTO-ing. This portion of the day specifically consists of reviewing his team’s code and pointing them in the right direction to optimize their work. The engineers and data scientists are so focused and truly putting their all into their projects, Dan said, to the point where it’s not just engineering, “it’s like an art.” 

Dan says his team is very receptive to that type of leadership—focusing on watering and blooming the flower that is the end goal with focused direction, instead of bogging them down with the weeds of details. They are focused on only what matters to them to get their job done, whether that job is working on an algorithm, designing an algorithm, designing a distributed system, designing a topological data analysis, designing a neural net, whatever it might be. “People should not feel like they’re at work. They should feel like they are living their life to the fullest.” That sentiment applies to Dan just as much as it applies to his team.

The second quarter of his day is focused on his projects. “As a start-up CTO, you are scrappy… you’re not only an executive, but you’re also an engineer or a scientist.” At the moment, he’s working on three projects in three areas of focus—a platform project, a data-science-engineering project, and a product-engineering project.

Dan says it’s tricky to turn on and off that individual contributor role because, in a start-up, engineers will “judge you just like any engineer.” There’s no leeway for mistakes just because he’s the head honcho. Dan has to be picky and realistic about the types of projects he works on so he doesn’t bottleneck anyone and so the integrity and quality of the work meets or exceeds the bar set by the rest of the team’s talent.

The third quarter of Dan’s job, as the coach of the ROOM8 tech team, is “watching film,” or the data science portion of his job. That means reviewing projects and making sure the technology is achieving a revenue goal. It’s also making sure the work is optimized and up to par with, or better than, industry standards. It also means participating in partnership meetings. It’s looking at the work produced with a neutral point of view, which Dan admits is hard to do. 

In the interview, the more we talked about the evaluation piece of his daily, CTO-ing routine, the more Dan shared about how important it is to self-edit, especially in a start-up. Self-editing as the CTO is vital for obvious reasons we’ve read about in this piece already. But self-editing as a professional, as an enthusiast, as a member of a community, as someone who’s striving to be better, it’s imperative to absorb what happens around you and act accordingly. That very skill could be the difference between settling for something good because it’s easier or cheaper or having the intuition and the rigor to achieve something spectacular—Dan and everyone at ROOM8, of course, strive for the latter.

The fourth quarter of Dan’s day is back to working on his projects. What’s different about the work from the second quarter and the work in the fourth quarter is everyone’s gone home—or everyone’s offline given the stay-at-home orders. He then has a day full of context and information that he now has a few hours to dedicate undivided attention to. 

What’s done in this time doesn’t *always* have to be directly related to ROOM8’s models and projects and such. But any time spent exploring any topic, Dan said, “I can only look at that if, and only if I can get a tangible outcome out of it,” admitting it’s hard to fight the inclination. 

Looking into a new programming language or strategizing a new way to think about their framework or exploring a new tool or reading a new report from McKinsey or keeping up with politics and privacy law are all things that interest Dan. But the integrity of his time is at stake. He must choose how to spend it wisely.

For Dan, a 12-hour workday is typical with six to seven hours of sleep and roughly four-five hours of free time to rest, which brings us back to the first and second rules of CTO-ing. He has an objective to increase the financial value of ROOM8’s technical asset at all times.

AI, Data Science, Machine Learning, ROOM8, Technology, co-living, entrepreneurship