Lilly Chen left high school at 16, spent about two years as a monk, and never earned a diploma. She still graduated from Colorado College in 2019, became an engineer at Facebook, and now runs FSH Technologies, a Philadelphia company that raised $25 million to build software for government paperwork. Her story is a case study in how a nontraditional founder turns an unusual background into a working company.
The practical lesson is not "become a monk." It is that Chen made a series of deliberate, copyable decisions: she picked a mission before a product, raised money pre-idea, and chose problems other founders ignore. Here is how each step worked, and what you can take from it. For related coverage, see Smart Home Automation: The Quiet Glue That Runs The Whole House.
How do you build a career without the standard credentials?
Chen's route shows that formal gates are negotiable when you can demonstrate the underlying skill. She was hospitalized repeatedly for Graves' disease, a thyroid condition, while on Medicaid. Her mother suggested they move to China and become monks. Chen was ordained in the United States, then spent months at a monastery in China with no internet access, meditating and rereading the books she had brought.
When her health became manageable, she returned to the U.S., homeschooled herself, and enrolled at Colorado College without a high school diploma. She turned 20 during her freshman year and graduated in 2019 with a double major in mathematics and economics. Along the way she played Dota 2 professionally and faced OpenAI's game-playing bots, an early brush with AI.
Then she did the unglamorous work: an initial job in Boston, then Facebook, where she built infrastructure used to train AI models. The takeaway for anyone without a conventional resume is concrete. Find one concrete skill, in her case programming, and get it into a real job before you worry about titles or credentials.
How do you decide when to quit a good job?
Chen quit Facebook shortly after Jan. 6, 2021, when she watched coverage of the U.S. Capitol attack during a remote company all-hands meeting. Her concerns about Facebook's role in enabling the organization of violence led her to conclude she could not work for Big Tech anymore.
Note what she did not do. She quit without a startup idea. She then worked through questions she had encountered as a monk: What made her happy? What mattered to her? What was she good at? Programming remained her most concrete skill. She needed to figure out where to apply it.
The actionable version: before you quit, know which skill you are carrying out the door. Chen's mission came first, the product came later, and that order held. This connects to our earlier piece, How To Set Up A Smart Home: Five Devices In Order.
Can you raise money before you have a pitch?
Yes, and Chen's route is instructive. An introduction through Envision, a student-run accelerator focused on nontraditional founders, connected her with Contrary, a venture firm that has backed Ramp and Anduril Industries. Contrary wrote her first check in 2021, before she had settled on the business that became FSH.
She had no developed company pitch. What won investors over, per Crunchbase News, was her earnest explanation of what motivated her and her passion for solving a meaningful problem. That gave her room to experiment. FSH took its current shape in 2024. The Contrary investor who first backed her, Jason Chen, later joined the company as chief operating officer.
The company has since raised $25 million in total, including a recent $20 million Series A that came together in about a week. Participants included Lachy Groom, General Catalyst, Ramp, Scribble Ventures, Operator Partners and Basis Set Ventures. Contrary also led the seed. The practical step here is the network one: apply to programs built for founders like you. Envision existed for nontraditional founders, and it produced the introduction that mattered.
How do you pick a problem worth years of your life?
Chen chose government software because, in her view, it could solve many consequential problems with little excitement attached. She had asked herself at Facebook: "What does it actually take to solve the really unsexy, boring stuff?"
Her diagnosis of why government processes are slow goes beyond an outdated interface. Information may sit in separate departments. Updating a website may require approvals from multiple stakeholders, including policymakers, communications staff and an IT team. A chatbot layered on top of that system, she argues, leaves the underlying problem unresolved.
That is a usable checklist for evaluating any problem space:
- Is the problem consequential for real people, even if it bores investors?
- Is the difficulty structural, in this case across departments and approvals, rather than cosmetic?
- Would a surface fix, like a chatbot, fail to solve it?
FSH applies this to some of the government's least glamorous tasks, including administering school food services and managing public records. Readers tracking the wider startups scene will notice the pattern: the least crowded problems often sit where the excitement is not.
What does this mean for non-technical founders?
Chen is the latest profile in a Crunchbase News series on founders from non-technical backgrounds who built venture-backed companies, alongside MagicSchool AI's Adeel Khan, Trunk Tools' Sarah Buchner, Tabs' Ali Hussain, Craft's Ilya Levtov, GC AI's Cecilia Ziniti, Incredible Health's Iman Abuzeid and Hilt's William "Will" Cielen. The series, and this Crunchbase News profile, suggests the pattern is common enough to study.
For more on how tools like the ones Chen builds fit into the wider software market, and how AI infrastructure work translates into products, our coverage follows the same thread: mechanism before takes.
The honest open question: Chen's monastery training shaped how she thinks about problems, down to a head monk's recurring question about whether she was the fish or the pond while feeding koi. The source material cuts off before explaining where that question led her. How her contemplative method maps onto product decisions inside FSH remains the part of the story we do not yet have.
