Diana Yuan is helping to lead the AI revolution. She is the COO of Boston’s scrappiest Machine Learning startup. Her, and her team of Olin Engineering dropouts, are democratizing machine learning tools for hackers all round the world.
Startup founders have an earned reputation for nonsensical over-confidence. Often, they’re impossible to spend time with due to a complete lack of self-awareness and social graces.
Diana is not one of those people.
As far as early-stage executives go, Diana has more self-awareness than most entrepreneurs twice her age. She’s bright and cheerful and a joy to be around. Having her on the show was a no-brainer for me.
This conversation was a lot of fun for me to produce. Thanks Diana :D.
This is the most important episode of Hacker Practice for aspiring non-technical startup founders. Diana Yuan’s role in the Indico origin story is instructive for any MBA-type/non-technical person looking for a technical partner. I have a feeling they’ll be telling her story in Babson business courses before long (if they aren’t already).
Enjoy this discussion on startups, fundraising, politics and technology with Indico’s Diana Yuan.
[4.30] Diana talks about the beginnings of Indico and beginning a machine learning startup
- Indico aims to revolutionize software through powerful, developer-friendly machine learning
[5.50] How Indico closed contracts before becoming Incorporated
[8.30] How Diana got involved in Indico by accident!
- Diana met Slater and Alec, Co-Founder of Indico by chance at the Affordable design and entrepreneurship class run by Olin College
[13.30] Being a non-technical co founder, Diana shares tips for joining forces and finding a technical partner
- Don’t force it
- Focus on the relationship you have with this person and ensure an aligned set of values and vision before getting caught up on technical capability
- Why your job as a non technical co founder is to translate
[19.30] Why Indico pivoted from a B2C to B2B (enterprise) model?
[25.00] Diana explains her experience of applying for and getting into TechStars
- Receiving funding from Rough Draft and having traction were helpful but networking was just as important
- Diana recommends networking and actively discussing your application with those reviewing it (e.g. Diana hit up the founders of TechStars, including Semyon Dukach, who was also a member of the infamous MIT Blackjack team).
[28.50] Diana shares her biggest challenges of being accepted into TechStars
[31.30] On being the sole college graduate amongst her co-founders
[35.00] Why the world is your classroom
[37.30] Why Diana identifies as an introvert and why it’s important to know yourself in order to avoid burnout
[40.15] Diana discusses the challenges of raising capital and the emotional stakes at play
- Don’t work with investors purely because they have money but because they will be a good fit for your team and offer a productive skillset
[46.40] Discusses the hiring process in a startup environment.
- Save employee time by putting technical filters up front in the hiring process to screen potential candidates
- It’s difficult to avoid hiring to keep up with growth but often startups have realisations and need to scale back. The important thing is to be conscious of it.
[53.50] The challenges of sponsoring candidates on a Visa
[56.50] Discussion on politics, America’s competitive advantage in entertainment and technology being eroded and technological libertarianism.
- Take away thought – “what happened 250 years ago is what is allowing today’s events to unfold”
[1.01.40] Diana discusses why transfer learning is Indico’s secret sauce
- Machine learning startups need to apply their expertise to specific market segments
- Transfer learning is a methodology for taking an approach for a specific problem, adopting it as a standard for a certain type of problem group and being able to apply it to new problems, that fit within that type of problem group.
[1.05.30] On maintaining an IP competitive advantage and staying up to date with cutting edge industry trends
[1.08.15] Diana discusses the kind of money required for cloud based infrastructure in a machine learning startup
- Diana discusses two innovative products offered by Indico, Crowd Label and Custom Collections, which help their end users label data and build custom models at speed and scale.
[1.14.00] How to protect IP and a business’s “secret sauce” in a highly competitive environment?
[1.16.45] The importance of choosing an investor and not feeling obliged to sign a term sheet just because one is put in front of you.
[1.17.30] What Indico does to prepare for a “black swan” event
- Risk scales in proportion to number of customers and stakeholders involved
[1.21.10] Why customers should come to Indico for machine learning expertise
- High quality unstructured text and video analysis. Free for first 10,000 API calls.
[1.22.50] ]How startups have capitalized on Indico’s unique product set at Hackathons to be able to build out MVPs in < 48hrs including a fake news detector
[1.25.00] You can find Diana on the Indico website because she runs the chat!
Subscribe to the podcast on iTunes or leave me feedback as a comment here.