From a nwHacks 2023 Honourable Mention to raising $1.65 Mil as a YC Combinator-backed Startup
Meet Reworkd AI, founded by Adam Watkins (CTO, SFU grad), Asim Shrestha (CEO, SFU grad), and Srijan Subedi (COO, UBC grad). Brought together by engineering classes and hackathons, the founders created a variety of different projects together before eventually forming Reworkd AI at nwHacks 2023.
The Origins of Reworkd AI
Reworkd AI strives to maximize the full potential of Large Language Models (LLM) such as GPT-4 by utilizing agentic behaviour. Their first in person working session was at nwHacks 2023, where they built a Chrome extension focused on personalized responses to digital communication. Now, it has evolved to allow users to configure their own AI agent that plans and completes tasks. The startup has now seen major success with over 26,000 stars on Github (at the time of publication) and is backed by Y Combinator and Panache Ventures.
nwPlus spoke to Reworkd’s founders to learn more about the vision behind their success and how their unexpected journey has led them to where they are today. Here’s what they had to share:
Could you share a turning point or specific moment when you realized that Reworkd AI had the potential to become more than just a hackathon project?
Adam: It actually started a couple months before nwHacks 2023. As a team, we were stuck in the ideation phase for a long time up until then. We came to nwHacks and said, “Let’s try something new”. GPT-3 was big at the time, so we thought we could make a Chrome extension and add some fun features. It was helpful for writing emails, and it did pretty well at nwHacks — so from there, we decided that this was something that we should pursue a bit more. Over the next 2–3 months, we all had full-time jobs, but it wasn’t until mid-April that [the project] really took off and we quit our jobs.
What were the key steps that took Reworkd AI from a nwHacks Chrome extension hackathon project to a Y Combinator-backed startup?
Asim: This extension was our first success as a company, our first indication that we were heading in the right direction with AI. Being a startup founder is just an exercise in doing micro pivots and adjusting to what you learn about the problem, your company’s strengths, and customers. We tried a lot of things before we landed on the AgentGPT idea. Adam and I hated waking up at 2 A.M. for on-call issues, so the idea was to work with AI agents (for operational events) acting as your copilot — if anything went wrong, it would tell you if a customer was having trouble with a line of code and how to fix it.
That’s what we were building at about the same time that the Y Combinator [application] deadline was coming up. There was a big change when we saw a huge market demand and popularity with this new idea of AI agents. Given that we were already working on AI agents, we did a micro pivot and built a low-friction general purpose tool that anyone could try. We launched it one week, and our lives essentially turned upside down.
Adam: I was traveling in Europe at the time, so I’d have to wake up every day at 2 A.M. because we were being bombarded by users. If our fingers weren’t on the keys 24/7, we would have failed to scale properly or gone bankrupt. It was a scary week.
Asim: We also suddenly had a lot of VC (venture capital) interest as well, and the intro to the Canadian VC firm that led our pre-seed was provided by a nwHacks judge (Chris Hobbs).
As you navigated the transition from a hackathon environment to the startup ecosystem, were there any surprises or unexpected challenges that you encountered? How did you overcome them?
Adam: The biggest thing that they don’t teach you in school is the business side to incorporating a startup. It gets incredibly complicated, especially if you want to incorporate in the States as a Canadian citizen to get your business up and running.
Srijan: We had to learn everything from scratch. We’d never incorporated anything before, so learning about where to do it, who to reach out to, and finding corporate legal counselors were all unknowns that we had to go through.
Adam: A startup is like doing a hackathon every single day. If you could imagine doing nwHacks for three months straight… it wouldn’t end well. We had to do a lot of self-regulation because you have to sleep in order to move faster — which took us a while to learn.
Asim: It seems that there’s a direct correlation between success and how much you can keep the hackathon environment going. I think we did this a lot better early on by trying to ensure that we had cool features out every week to share, following a hackathon-like procedure in trying to make things as fast as possible.
How did the experience of developing a project at a hackathon contribute to your understanding of entrepreneurship and building a startup?
Adam: Building a successful business is making sure you’re basically doing the hackathon every day. Especially in the early days, hackathons are a great example where you may not need to focus on code quality and you want to cut corners to deliver value as fast as possible. A startup is basically the same thing; it’s not until you grow into an enterprise that you have the luxury or time to go back and do things the right way.
Srijan: Moving quickly is the biggest advantage you could have as a startup itself, e.g. the ability to create an MVP (minimum viable product) within five weeks.
Asim: A word of advice — keep the essence of having fun as you’re building a company. Being a startup is a five or ten-year process, so you should at least enjoy it as much as you can.
Looking back, what do you believe were the key factors that set Reworkd AI apart from other projects and contributed to its success as a startup?
Adam: Being open source helps you early on as a startup because it encourages people to engage with you. It’s critical to your initial startup journey and it’s interesting how it loops back to the hackathon experience.
Asim: Having worked on this together for ten months, we had an understanding of what was lacking in the market and what people were excited about. We were able to make this project during the inception of this modern idea of AI agents and really capture the attention and value that people were interested in.
Srijan: The project that we were working on beforehand was more catered towards on-call processes like AI agents. Within a week or so, we took the parts from that and made what AgentGPT is today. During that time, we moved fast — we knew the condition of the market and what was needed, so we made an MVP within a week.
Adam: Another factor that sets us apart is that you have to make sure the founders vibe together. Hackathons are a really good way of doing a “vibe check” amongst your team and making sure you work well in a high-stress environment. We see among some founders that there’s a correlation between how well your founders get along and how successful your project is long-term. It’s really important that everyone has a shared attitude and vision. Try to do hackathons with a new team every time. You learn so much more as opposed to being in an echo chamber. By having that variety, you’ll learn naturally what makes a good coworker, founder, or roommate.
What aspects of embarking on this startup journey as first-time founders have you found to be the most fulfilling? Most challenging?
Asim: I was essentially trying to run a startup even when I was at Amazon. I spent a lot of my time working on a side project, worrying about marketing and distribution, and talking to customers. And suddenly I had an extension that 1000 engineers were using in Amazon and now I can do all the best parts of that as my full-time job. I think that’s incredibly rewarding and incredibly exciting to wake up to. It was challenging learning that I didn’t really have the business acumen side and trying to think through things like going to market, churn, and ideal customer profile (ICP) that I wasn’t familiar with before.
Srijan: The most challenging for me would be the unknown, which is also the fun part: navigating through that and trying to find out which product and which features work because you don’t really know until you test it out. It’s both challenging and very fun, and when you get it right, it’s very fulfilling.
Adam: I think the most fulfilling thing is you get to work on the technology that you enjoy the most often — you’re working right at the edge. Prior to this, I did Enterprise Java and some of the features I was writing were either boring or had very little impact. Working on your own business ensures that you’re trying to have as much fun as possible for years to come. In terms of challenges, when you’re very invested in something, it becomes very emotionally draining over long periods of time. It gets challenging to keep yourself in a headspace where you can keep going for ten years down the road.
What is the biggest piece of advice you have for future hackers at nwHacks 2024 who want to evolve their hackathon project idea into a successful startup?
Adam: First, be persistent. If you want something to be successful, you have to put in the work. If you do the hackathon one day and then you forget about your project afterward, chances are it probably won’t be successful. But actually putting in the work day over day dramatically increases the odds that something will come out of it.
Second, don’t set your hackathon project in stone as something that will be your business long term. Between our hackathon project and where we are now, we’ve probably pivoted five or six times. You have to make sure you’re comfortable doing new projects every two weeks until you find the right one; even now, I don’t think we’re quite there yet.
Asim: At least in Vancouver, the end goal of university students is always to get into big tech and big firms. There’s just so much more of a rewarding journey going through the startup path and I definitely encourage everyone to at least try. The biggest thing that you can do is to work on cool things and get it out there. Our story can essentially be boiled down to: we thought of some cool idea, we launched it on Twitter, and we gained a lot of traction. It’s definitely something that could be emulated by everyone else that’s going to nwHacks. I’m also happy to give advice or if you want to reach out to us and talk through what it’s like to be a startup.
Srijan: If you want to move your hackathon project into more of a startup, you have to do something you enjoy. Realize that your idea of doing the hackathon might not be the actual startup itself, but if you’re in that particular space that you actually enjoy — for us, that was AI — then you will always enjoy building the project out. If you push hard enough, then one day it will probably work out.
Adam: It’s also important to be realistic and honest with yourself. If you do want to go down the startup path, it probably means you won’t have weekends anymore. But you have to learn to have fun during those times, and hopefully code on your weekends anyways.
Learn more about Reworkd AI or reach out to them at their socials below!
Website | LinkedIn | Twitter / X | Instagram
Written by Jennifer Shui, Content Writer
Edited by Victoria Lim, Marketing Director
