- Startup Life: Unscripted
- Posts
- Startup Life: Unscripted #41 with Alex Chin, AI Engineer at Motif Analytics
Startup Life: Unscripted #41 with Alex Chin, AI Engineer at Motif Analytics
From Stanford to startup, Alex reflects on adapting academic skills to lead data science innovations at Motif Analytics.
Startup Life: Unscripted is a TNG Media newsletter, as part of The Nudge Group, where we feature candid conversations with startup operators about their career journeys and experiences. If you received this email as a forward, you can read all our past interviews and subscribe right here.
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/6a10a01d-768f-4865-bb5d-1816ff0ae3ec/beehiiv-banner.jpg)
Welcome back to Startup Life: Unscripted! In today's edition, we're excited to welcome Alex Chin, who is currently steering the data science initiatives at Motif Analytics. He talks about the experiences that shaped his career trajectory, offering insights into his transition from statistical theory to practical application, and how these experiences have culminated in his role at Motif Analytics.
Key interview takeaways:
🔧 Theory to Application: Alex discusses his journey from being captivated by statistics to applying rigorous methods in diverse domains. At Motif, he's leveraging his expertise to reimagine product analytics, enhancing how data teams make strategic decisions.
🌐 Academia to Industry Dynamics: Reflecting on his shift from Stanford to the startup scene, Alex highlights the adaptability required to thrive in both environments.
🛠️ Experience Across the Spectrum: From a data consulting startup to a tech giant like Lyft, Alex contrasts the cultural and operational differences he's encountered and how these experiences have enriched his approach to data science and team leadership.
📈 Innovating with AI: Looking forward, Alex sheds light on integrating Large Language Models into product analytics at Motif. He discusses the potential impacts of AI on the industry, focusing on innovative applications that go beyond traditional data analysis techniques.
Got a minute? Share this newsletter with your network and help us inspire more individuals with real-life startup stories and valuable insights.
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/a8343a32-871f-4bbc-b0ca-d53e64e741b9/Startup_Life_banner__1_.png?t=1713334734)
Alex, with a rich background in data science and AI, could you share what sparked your initial interest in these fields and how it led you to Motif Analytics?
I was drawn to statistics as a field because of the intermingling of theory and application: developing statistically rigorous methods and then applying them to diverse domain areas.
Once I started working, this interest grew into developing and building tooling for practitioners who work with data. Motif is tackling this on a grand scale by reimagining the field of product analytics and how data teams make decisions.
I happen to enjoy a healthy mix of software engineering, methodology research, and empirical data analysis — At Motif I’m able to work on all of these things because they are all crucial to the success of the product we are building.
Transitioning from academia to the tech industry can be a significant shift. How did your time at Stanford, particularly working on statistics and causal inference, prepare you for the challenges in the startup environment?
They are certainly very different environments! One of the key lessons I learned in grad school was how to embrace and navigate ambiguity effectively.
Academic research is inherently exploratory: it starts with a general direction but the specific path forward—the right questions to ask, the optimal allocation of time and resources—are quite undefined. In this environment it’s important to be flexible, to pursue multiple ideas at once, and to be comfortable if you don’t get any concrete feedback for months or years.
I’ve found working at a startup to be similar in many ways. We have a vision of the product we want to build but the landscape changes rapidly and we need to adapt quickly and pivot strategies when necessary.
I also did a lot of interdisciplinary work during my PhD and got in the habit of reading literature from different fields like computer science, economics, and the social sciences. Often people are working on very similar things but have a slightly different approach or use different terminology.
The breadth and exposure to different perspectives can be especially helpful when building 0-to-1 products at a startup, in contrast to working on a more mature product at an established organisation.
You've had the opportunity to work at both a data consulting startup and a major player like Lyft. What differences did you notice in the work culture and approach to data science between these two experiences?
One of the biggest factors of working in a consulting environment is that you are working in collaboration with the client on their projects using their datasets.
This can create dependencies on external factors outside your control, such as friction in accessing data. On the flip side, it’s really useful to see the kinds of problems that people care about and the solutions they want, and I appreciated being able to work on a variety of projects across different industries.
At Lyft I always had access to data as I needed it for any analysis I was trying to run. The objectives of what we were trying to accomplish and optimise were clearer because they are internally defined. It can be more difficult, however, to make large, impactful changes due to the procedures and governance that helps a large, complex organisation run smoothly.
I worked on an internal research team at Lyft, which meant I was often consulting for the various product teams within Lyft and in many ways gave me the best of both worlds. My advice to folks early in their career is to try experiencing different types of companies and roles. It’s a good way to figure out what you want to do and what types of environments you thrive in.
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/72a0fb00-9d5b-46e4-ac27-7ef8d2f91704/fundabl-2.png?t=1697622884)
🎗️ A quick word from Fundabl 🎗️
Looking for growth capital, a bridge towards your next capital raise or year-end funding round?
Be sure to have a chat to the guys from Fundabl. They offer founder-friendly funding for start-ups, scale-ups and high-growth businesses.
You’ll never miss an opportunity again, with up to $3m of non-dilutive, flexible funding and a quick & easy process.
P.S. We can vouch for the experience! 🙌 🙌
At Motif Analytics, you're bringing cutting-edge Large Language Models (LLMs) into product analytics. Can you shed some light on what that involves and the potential impact it could have on the industry?
LLMs have become very good at translating natural language into code because they have been trained on high-quality SQL and Python tutorials that are available on the Internet. The resulting copilot tools can increase productivity for analysts when you know the exact data you want to pull and the charts you want to generate.
However, I find the bigger challenge is not just in translating queries to code, but in being able to pose complex questions in the first place. What are the key factors driving the outcome I care about? How should I change the levers I control to optimise a particular outcome? Can my dataset actually answer the questions I have? What additional insights are lurking within the data that haven't yet been considered? LLM products like ChatGPT often struggle with these questions because they’re difficult to answer using SQL and existing data tools.
At Motif, our focus on sequence analytics and event log data means we are uniquely positioned to leverage the same sequence-based deep learning models that power LLMs but trained directly on the datasets that analysts care about.
Just as language models pick out patterns in historical text in order to generate the next word, Motif’s AI models can mine event log data to understand the patterns driving future outcomes. The use of AI in this way is quite different from the analysis chatbots that are more common, and I think could be quite transformative for the industry.
What's a typical day for you look like? Can you take us through how you balance hands-on technical work with other responsibilities?
I’m someone who loves to have a lot of different projects on my plate at once (sometimes to my detriment) and I’m not sure there’s a typical day for me!
I spend a good amount of time training, evaluating, and optimising the architectures of our core machine learning models, as well as designing and prototyping the front-end visualisations that go into the end product.
At Motif we take pride in exposing rich, interactive visualisations to the end user and this holds true for our AI model outputs as well.
I also help out with writing production code for our app where I can, and doing data analysis for our customers. More recently I’ve been focused on building out our AI infrastructure to make sure that we can reliably handle enterprise datasets that comprise billions of data points.
In your view, what are the most exciting trends or potential breakthroughs in AI and data analytics that we should be looking out for?
I think multimodal systems are really interesting and I’m excited to see them continue to mature! Beyond that, LLM applications have primarily focused on text, images, audio, and video — I’m interested to see how they can be applied in domains like causal inference and experimentation and other typical areas of statistical analysis. We’ve started to do this at Motif and I think the potential is sky-high in this area.
Lastly, for those aspiring to enter the AI and data science field within the startup ecosystem, what advice would you give? Are there specific skills or mindsets they should cultivate?
Becoming comfortable with ambiguity is a really useful skill, as I mentioned above. I’d also say it’s helpful if you can get good at learning and adapting to new tools, especially in a world where the AI landscape is changing so quickly. (I use the word “tool” broadly here — it could be a software package, algorithm, model, SaaS platform, etc.)
I’ll hear about something new on Twitter or LinkedIn or arXiv and am happy to play around with it for a few hours. Many of the software packages and platforms that I use on a daily basis today are ones that I hadn’t heard of even 6-12 months ago, but now I can’t live without them.
From the Startup Life team
And that's a wrap! We hope you've enjoyed this edition as much as we loved putting it together. Stay curious, keep learning, and above all, enjoy the rollercoaster ride that is Startup Life. Catch you in the next one! 👋 Not subscribed yet? Do it here and don't miss out! Subscribe Now.
🎙️ Check out Give It A Nudge, our video podcast that explores the world of start-ups, founders, venture capital, and private equity. Follow us on YouTube, Spotify, Apple Podcasts (or your preferred platform) to hear new episodes every fortnight.
📩 Want to get your brand in front of 14,000+ startup enthusiasts, operators, and potential candidates? For details on our sponsorship opportunities, simply hit reply and let’s chat!