Research & Analysis

Nashad Singh Described FTX’s Founding, Early Days at Alameda on Podcast in 2020

FTX’s shadowy co-founder Gary Wang has not, to my knowledge, ever given a public interview. However, another key player who worked with him has: Nashad Singh.

Singh is often characterized as the fourth member of the “inner circle” of the FTX/Alameda operation, along with Bankman-Fried, Wang, and Alameda CEO Caroline Ellison. After FTX collapsed, he was named by the Wall Street Journal as having been “aware of the decision to send customer funds to Alameda,” along with the other three.

An understudy of Wang from the moment he joined Alameda, Singh is the primary developer who is credited with helping Wang write the code for FTX, where he (Singh) was given the title “Head of Engineering.”

Screenshot from the now-deleted “about” page of FTX’s website (archived here)

In December of 2020, Singh gave an interview to The FTX Podcast, hosted by Tristan Yver. In it, he provides a unique insider’s perspective on the murky early days and inner workings of Alameda and FTX. Perhaps most notably, this includes insight into Wang’s role at these companies — something that we are yet to hear from Wang himself.

Here are some of the more interesting takeaways from the episode.

1. Singh didn’t get involved with Alamenda until about a month after it was founded, and at that time he estimates that there were already probably around “five people in the apartment.”

YVER: And going back to Alameda, I want to hear about your experience. So, from my understanding you went and visited Sam’s startup. What was going on there? Where were they? How many people were there?

SINGH: It was in an apartment at the time; it was quite early. I think I first visited Alameda when it was about like a month into– like into its existence. And, I think there might have been like five people in the apartment. But, you know, at that time it was really chaotic, it was unclear who was there part-time, who was there full-time.

And really, what was obvious was that the most important thing— things they wanted to do, were really important and really fruitful. I sort of watched Sam, like, execute a sequence of trades. I knew nothing about trading at the time, but even then it was sort of understandable that sequence of trades was super profitable, and easy to understand, and that there were lots of them available. And so, there was all of this sort of operational flurry around getting Alameda to a state where it could take advantage of these trades. It was really fun, it was really chaotic, and it was clear that it could possibly be onto something pretty big.

Timestamp: 2:52

2. Singh worked part time for Alameda for about a month before quitting Facebook to go full time.

YVER: And when you started, how was that first transition phase? I’m guessing it wasn’t that you just one day to another left Facebook — did you test the waters a little bit? How did that go?

SINGH: Yeah, so I spent, let’s see, maybe about a month doing weekends and nights at Alameda. So I’d like, ya know, have my day job in sort of South Bay in California, and then I’d drive up to Berkley, and go work in the apartment or whatever. And at some point it became obvious that it was like kind of stupid that I was even bothering with the day job. And so I like took some time off, and really gave my 100% [to] working at Alameda. And then—it was obvious even earlier than this, but it took me some time to really arrive at it myself—eventually decided to make the switch. But it was about a month of toying with and actively working on Alameda before I formally made the switch.

Timestamp: 9:55

3. Singh cites effective altruism as his motivation for leaving his job at Facebook for Alameda. (Note: the discussion about EA continues for several more minutes after the part transcribed below)

YVER: And I imagine there was more to it than that, though, for you to decide to switch, right? Cause I believe you have other incentives. Would you mind breaking that down a little bit?

SINGH: Yeah, um— is my interest in effective altruism. EA, effective altruism, is this idea that you can do a lot of good for the world if you’re careful about choosing how you do so. And that it’s worthwhile spending your time thinking about what the most effective ways to do good are. And that the most effective ways are sort of significantly better than what would happen if you sort of went about it haphazardly. It’s a really sort of big, hard, lofty goal to improve the world a lot. But lots of smart people have thought about this and made a lot of headway, [and] found really sort of straight-forward, exploitable [sort of[ opportunities. And then of course there’s sort of other ones that are riskier and more uncertain, but for which the upside is possibly huge. As an example of something that’s sort of obviously really a good use of time, money, etc, uh, is getting, like, malaria bed nets. You know, for something like nine or ten thousand dollars, you can, in practice, save, like, a life’s worth of happiness or something. They’re– definitely getting fuzzy with the definitions here, but, you know, something to that effect. And that’s super encouraging…

Timestamp: 3:44

4. When Singh first arrived at Alameda, his ability to add to the code base (“invent”) was limited, and he was mentored by Wang:

YVER: I’d love to hear a little bit about your experience and your growth as an engineer during this time at Alameda. Because I imagine it must be a very self– I mean I know that it’s a very self-starter culture. So I’d like to know how you navigated that.

SINGH: Yeah, you’re right. It was really self-starter oriented. Like, it did take, you know— one thing that was nice about joining Alameda early is that there wasn’t much of a code base, and so there was something of, like, you know– I could sort of invent whatever I wanted. But at the same time I didn’t really have the ability at that time to do invention really well. Gary, our CTO, is sort of brilliant beyond belief, and was like a really good mentor. And I think being forced to push things out on really sort of short time scales, and design thoughtfully, was a really good forcing function for me to level up.

Timestamp: 10:35

5. Even after getting to a point where he could contribute “uniquely” consistently, he wasn’t “the expert,” and remained an understudy of Wang:

SINGH: I think, in the beginning, ya know, it was challenging. I’d see like a list of things that we wanted to get done. I had some idea of how to do each of them, but for every single one there was something in the way. And I think, you know, the process looked something like: find a task that people want done; think about what I know about how to do it; bug Sam and Gary, hopefully for not too long, on the things that are sort of blocking me on it. And then— and just try to chip away at it. And pretty soon, I definitely wasn’t like the expert on how to build stuff, but I was comfortable enough in our systems that I could really uniquely contribute, and really consistently do so. And then you sort of get, you know, a lot of like exponential growth, and develop a lot of competency within the context of like a codebase for an organization. It was fortunate that, like, Sam and Gary were there to give a lot of really great really quick advice, and also that the codebase was up to us to develop, and I didn’t have to sort of, you know, learn a billion company-specific conventions. That was for us to make up.

6. After they decided to found FTX, Gary and Sam “took off to Hong Kong” and built “the first version” of the exchange “in a couple month” without Singh, who stayed in Berkley managing engineering for Alameda during this period. (Note: “MVP” stands for minimum viable product, “a product with enough features to attract early-adopter customers and validate a product idea early in the product development cycle.”)

SINGH: …and so I think for like a year we were like, obviously exchanges can be a lot better than they’re doing. And so there’s like, huge room for improvement here. But we were kind of afriad to do it ourselves because of unknown unknowns. And, I think we just sort of took the math on the upside seriously, and the likelihood that we’d get there seriously, finally, a year after we should have. We’re like “Ok, we should do this.”

Gary and Sam basically took off to Hong Kong—uh, Gary being our CTO—and built the exchange in a couple months. Like, the— the first version of it. In that period I— this is another level up moment for me— I basically took over Alameda engineering, and was managing the engineering team and sort of leading, uh, like all engineering-related efforts back at home base in Berkley.

When the FTX MVP was sort of done, and when one guy under me, Nate, who’s really great, was sort of ready to take the reins at Alameda, we did like a little swap, where I went to go work, you know, I sort of stopped working Alameda, and haven’t that much since at all, and went to go work full time at FTX. And Nate took over for Alameda

INTERVIEWER: Did you fly to Hong Kong as well then?

SINGH: Yeah, I think I flew— I think, I forget exactly when. It might not have been like exactly the same time that I started working at FTX, but soon after.

Timestamp: 16:13