Life
Opendoor Experience
Near the end of my time at Quora, I got a Facebook message from JD Ross saying he wanted to meet me in SF, on a recommendation from Lucy Guo. I went up to meet him, and it turned out he was a founder at a startup trying to do something different in real estate: they wanted to buy and sell houses as a company, and digitize the whole home transaction process.
First Impressions
They showed me around the office. I talked to engineers, saw the internal tooling they had built to start automating the process, and I was basically sold. The place was dynamic, energetic, exciting. I wanted to be there.
Before I signed the offer letter, JD and Eric Wu (the CEO) offered to take me out for dinner somewhere in San Francisco. I didn't know where to go or what to eat, so I just randomly searched up something I thought wouldn't make me look like a 23-year-old idiot with no clue about good food. We went to La Briciola.
I can't remember anything we talked about, probably because it was a pretty chill night with steaks and fine wine (at the time, and even now, I barely know anything about wine, but I'm sure they bought some good stuff). Eric intimidated me. Whenever he looked at or talked to me, I got this sense that his mind was processing about five things about me all at the same time. It felt like he could read my internal thoughts at the same time I felt them. That was strange to me.
Years later, looking back, I think I can do that to some people now — but I'm not sure they can tell I'm doing it, the way I could tell with Eric.
Diving Into The Work
I joined the company — it's called Opendoor. Work was fun in a way it had never been before. On one of my first days (it may actually have been my first day), Daniel LaGrotta started pair programming with me to get me making changes to their internal tool, which we called the admin tool. That was how Opendoor ran: there was a consumer side — a beautiful website and landing page where people could request offers — and behind the scenes the whole process was encoded into workflows that operators at the company managed through the admin tool.
Faking Scale With Humans
Mostly everything was done by the operations team. Manually, using the admin tool. This was a lesson that hadn't even crossed my mind before: you could fake scale by having humans do the most difficult parts of the process, the parts that were hard to automate.
It was a revelation. I had always assumed technology should handle every step of a process. Opendoor took a more practical view: use software to make humans more effective, and worry about replacing them later — especially in the early days of trying to disrupt an industry.
Building for Operations
I quickly realized the most valuable software I could build wasn't for consumers. It was for our operations team. Every efficiency gain in the admin tool multiplied across the organization — save each operator ten minutes a day, across dozens of people, and it adds up fast.
That meant getting extremely close to the operations teams: sitting with them, understanding their workflows, finding the friction. I'd shadow someone for half a day, then go straight back to my desk and ship changes to fix what I'd just watched them struggle with.
More Than a Software Company
What fascinated me most was the business model. Opendoor wasn't just building better technology for real estate — it was becoming a market maker for homes. That took pricing algorithms to value houses accurately, operations workflows to inspect and repair them, capital markets expertise to finance the purchases, real knowledge of local markets to expand geographically, and an understanding of what would make a person comfortable selling their home to a company at all.
Technology was just one piece of the puzzle, which was refreshing. I was constantly learning from experts in fields that had nothing to do with software.
The People at the Top
The leadership team impressed me in ways I hadn't experienced before. Eric could dive deep into the pricing models one minute and lay out a vision for transforming real estate the next. JD, who had recruited me, had an uncanny knack for spotting consumer pain points and turning them into product. Ian Wong, the CTO, showed me what deep technical skill looked like when pointed at an industry as traditional as real estate. Being around that caliber of people early in my career changed how I approach problems, and I've carried it with me since.
Automating in the Right Order
The other lesson was about sequencing. We didn't try to automate everything at once. We automated the highest-leverage parts of the process first, built tools that made operators dramatically more efficient, collected data from the decisions humans were making, and then gradually automated more as we understood the edge cases. The company scaled much faster this way than if we'd tried to build fully automated systems from day one.
Reflection
Opendoor shaped how I think about technology in traditional industries. The interesting work happened at the intersection of software, operations, and business model — no one piece carried it alone.
More than anything, Opendoor showed me that software doesn't have to replace people to be transformative. Sometimes it just has to make them faster. That idea has informed nearly everything I've built since.