amber huang
I’ll start by answering the most-asked question for those that only look at where I began and where I am now. The answer is no. Obviously, no. My background has nothing to do with what I do now. It’s ridiculous to question whether packaging design has any use in computational drug discovery. In fact, if it makes things easier, you can just think that I never went to college and don’t have a degree at all.
I am a self-taught data scientist and one of the founders of SieveStack, an AI drug discovery company. I’m writing my story because I want to meet others who have been through a similar journey as mine; people who came from a non-technical background and worked their way into deeptech. In today’s Silicon Valley, this has proved to be difficult.
I’ve always loved design. I have a naturally good eye for it, and to me, that offered enough of a scaffold for a child to cling to while trying to figure out who she wanted to become. In college, I majored in design, because it was the skill that I was the most confident in making a stable career out of. As I completed more courses for general requirements and my minor, Sustainable Environments, I realized that there were areas of study out there that might be much more interesting to me. I thought often about changing my major, but always dismissed it. More than anything at the time, I was concerned with graduating as soon as possible so I could become financially independent from my parents.
Over time, I became awe-inspired by the kinds of changes being made by political activists, scientists, engineers, and others working to shape systems and tackle large-scale problems. But I thought to myself that those occupations are for people more capable than me, and out of them only the smallest fraction of the top are making the real changes in the world. I convinced myself it wasn’t a path I could take—that I didn’t have what it takes.
In the summer of 2020, I took a job in Philadelphia to work for a toy licensing company. I was excited to design packaging for a brand I’ve loved since I was a child: Pokémon. I adapted to the role quickly and was promoted early on, but over time, the work began to feel repetitive. The reality of designing for a well established brand is one of limited creativity and rigid brand guidelines. After a year and half, the company mandated a return to in-person work post-COVID; I was not having it. COVID had spoiled me—if I was going to do boring work, it should at least be remote.
In Spring 2022, I started working for an apparel licensing company and was designing packaging for socks. Yes, socks. Somebody has to do it. I knew that it would be a mediocre job, but it was fully remote and I thought that I wanted the digital nomad lifestyle. Turns out, I could not deal with the boredom for 40 hours/week. So by the end of 2022, I quit; I did not wait until I had a backup plan or another job lined up.
I couldn’t envision myself doing similar types of work any longer. Sure, there will always be room to improve my graphic or structural design skills; there are ingenuities to be made about how to cleverly fold and cut a piece of paperboard to make the most efficient package in the most cost effective way. I loved all that. I really did. But these problems were like fun puzzles and they weren’t meaningful enough for me in the grand scheme of things. If I was going to spend the most of my waking hours doing a certain thing, it better be worth my time.
A friend I was traveling with at the time offered to teach me how to code in Python. I was hesitant at first because truthfully, I didn’t want my friend to think that I was stupid if I ended up being bad at it. They convinced me after we discussed all sorts of new career paths that could open up if I knew how to code. I began with the goal of eventually working in climate tech or at a climate-focused nonprofit as a data analyst—an area I’d been passionate about for a long time.
I was lucky to have my friend who helped me get over the initial learning curve. Even so, it wasn't all fun and games.
I still remember learning the difference between a for loop and a while loop from w3school, then going on to take my first Coursera course, “Programming for Everybody” from the University of Michigan. I have fond memories of learning about the Fibonacci sequence from my friends during new years eve while we ate ice cream cake; I also have bitter memories of breaking down in my room, feeling utterly incompetent and defeated after days of not being able to solve a double recursion problem within optimal time and memory constraints. As I started taking more specialized courses in data science and machine learning, the more glaring my education gap became. There was a moment when I realized that I don’t remember anything from high school calculus so I have to go back to pre-calculus first, just to then realize that I also need to relearn trigonometry first.
There were many times when I doubted myself, feeling like I wasn’t “smart enough” to keep going. But each time I overcame a challenge, it reminded me that setbacks are temporary. What truly kept me going—even on my worst days—was the conviction that leaving my old career was the right decision; that I rather grind through another course for a chance to work towards something more meaningful or even just more intellectually stimulating than to change the color of some text by increasing two percent of magenta on a package that no one will spend more than two seconds looking at just because my boss told me to.
Other than my insecurities that initially held me back, the hardest part really wasn’t the material itself—it was figuring out how and what to learn, because I didn't know what I didn’t know. There was an instance that perfectly captured this sentiment: upon asking my friend (who is a mathematician) for his intuition about a problem that had stumped me—“find the nth Fibonacci number modulo m”, he immediately noted that since there are only so many remainders modulo m, they would probably start cycling. I asked him how he knew this, he replied that it was just something he had “always known”. I was bewildered at his answer. It felt unfair. How could I have known that if I had no prior knowledge about Fibonacci numbers?
I started applying for data analyst jobs around Spring 2023. The timing could not have been worse, as there were massive layoffs happening in the tech industry. Over the course of two months, I received one rejection email from a climate nonprofit and no replies from the rest. I’ve been without a job for several months at this point and needed income.
I don’t quite remember exactly how I got the idea, but I was already fairly familiar with the ecommerce reselling market for fashion as a buyer and casual seller of my own clothings, so I decided to try and make some income from being a reseller, but this time, leveraging data. I knew that full-time resellers usually play the numbers game and buy liquidation pallets or wholesale, but I hypothesized that with insights from data, the acquisition of inventory could be done with more precision. The main idea was quality over quantity. My plan was to scrape as many old listings as I can from eBay and other reselling platforms, then train machine learning models that predict which active listings are underpriced with high potential of profit margin and turnover time for resell.
In short, it worked. For the rest of 2023, I continued to refine my models while taking more courses on the side to increase my chance of getting hired the next time I apply for jobs.
In the beginning of 2024, I decided to wind down the reselling business. It may sound naive, but I still wanted to work for a larger-than-life cause. Maybe it’s because I’d gotten used to working for myself, but the idea of working under someone else was just unappetizing. With all this in mind, I started looking into how to launch a nonprofit organization.
As I looked deeper into how nonprofits work, I learned that (1) it’s actually a problem when there are too many nonprofit organizations in the same space; and (2) unless the smaller organizations have some insight into making a real change to the cause, they end up slowing down the overall movement by taking up resources that could’ve been allocated to more established organizations. So, I concluded that I did not want to start one until I had more knowledge and a concrete plan.
Around the same time, my friends Ian and Slava told me that they were planning on starting a company. Since I was again jobless, and seeing how busy my friends were, I offered to help them with their web app. They were making a web platform for molecular modeling, a concept which I was completely unfamiliar with. And I didn’t really need to understand it since I was just coding up the interface, but naturally I was curious about the subject that compelled my friends to start a company around. They patiently answered all my questions—though I lacked the biology background to fully understand some of these answers, while others I was able to grasp from a technical or logical perspective.
I was overall intrigued, and it was perhaps opportunistic, but I wanted to get more involved. After some discussion, my friends tasked me with building a post-processing algorithm to analyze top-performing chemical compounds based on their early models. Then, with a need to understand exactly what I’m spending my time on, I started doing market analysis and began to form my own opinions on the different technological methods in drug discovery, including the ones that my friends were working on. Soon, my understanding of the field was on par with theirs and my input on strategic decisions began to carry more weight.
I suppose I impressed Ian and Slava enough that they asked me to join the company. At this point in the story, it feels pretty unsatisfying that I didn’t end up doing anything related to climate or nonprofit, doesn’t it? The writing for my thoughts on why I ultimately decided to commit myself to the field of drug discovery will have to come another time.
In April 2024, Ian, Slava, and I incorporated the company and officially became co-founders. Slava wrote about the inception of SieveStack.
I dedicate this essay to you. A friendship like ours is rare, and I will always cherish it. Game recognize game. Thank you.