Lucas Ewing is CEO and Cofounder of Lilac, an open-source tool ensuring that data scientists always have enough GPUs for their work.
Lilac takes scattered compute from anywhere and connects it into one single, unified compute fabric for managing training jobs.
Lucas and his brother Ryan cofounded Lilac with a vision of becoming the foundational layer between data scientists and infrastructure.
Lilac is currently just a 2-person company and part of Y Combinator’s Summer 2025 batch. Check out Lilac’s launch video posted on YC’s LinkedIn page.
Early Life: Minecraft Modding in Middle School
Lucas grew up in Sweden, but his father is American and eventually brought the family back to the U.S. They settled down in Concord, MA and Lucas attended The Fenn School.
Lucas’s first technical experience was teaching himself Java for Minecraft modding. At the young age of 13, Lucas understood “make something people want.” He built the 2nd Herobrine mod ever created at the time, that got 50k downloads, and people made YouTube videos talking about his work. (Check out luew12345 on Planet Minecraft.)
High School: Data Science Competitions and AI Research at Harvard
In high school, Lucas gained an interest in data science and he started the “Big Data Club” with another student at Concord Academy. In 2018, they wanted to expand the club beyond their school, working with other local high schools like Concord Carlisle High School.
Together, they found a competition hosted by Teradata, the data warehouse system, from Tableau, the analytics tool for business intelligence. The data analytics challenge was part of the Teradata University Network, intended for students at the collegiate level, but they applied anyway as a team of high school students and were accepted.
Their project ended up winning the competition’s People's Choice Award for their project submission “Guns in Schools” on the impact of declining mental health in the US and gun violence (Teradata’s 2018 Analytics Challenge Winners). Winning the award at the Teradata Analytics Challenge was a key moment in Lucas’s data science career journey.
After winning that award, Lucas’s high school was very proud of his achievement and this opened up opportunities for Lucas. It led him to a summer internship as a Research Fellow at Harvard University working with a professor Nobuhiko Hata, PhD in the Surgical Planning Laboratory at Harvard Medical School for an AI applications research project in collaboration with the teaching hospital Brigham and Women's Hospital in Boston, MA.
Through this research experience, Lucas transitioned from data science to application using AI. He used a random force classifier to do automatic segmentation of tumors in prostates and published his work in the SPIE (Society of Photo-Optical Instrumentation Engineers) JMI (Journal of Medical Imaging) as first author of the paper, “Pixelwise tissue segmentation for precise local in-vivo dose response assessment in patient-derived xenografts” (https://doi.org/10.1117/12.2513080) March 2019.
College: Balancing Tech and Medicine Interests During the COVID-19 Pandemic
After high school, Lucas went off to Harvey Mudd College in Claremont, CA to study Computer Science in 2019. In addition to loving tech, Lucas was interested in medicine and decided to take pre-med course requirements as well.
He became an EMT with Trinity EMS while home in Massachusetts during winter break in 2020, just before the spring semester of his freshman year when COVID happened. He then suited up in a biohazard suit and played an active role helping patients throughout the pandemic.
During his junior year of college, Lucas joined Nara Logics, a Series A startup working on an open box AI. He worked as a software engineer for Nara Logics throughout his last year of undergrad, while also working as a student engineer for Shopify on his senior capstone project.
When Shopify’s CTO Allan Leinwand became the CTO of Webflow, he personally offered Lucas a role there. Lucas worked at Webflow after graduation, learning data engineering and deep tech in a professional environment for two years before starting his own company, Lilac.
Lilac: Starting Lilac and Getting Backed by Y Combinator
When Lucas realized that the current AI infrastructure will not be sufficient for deep tech needs in the future, he and his brother Ryan Ewing, who was previously at AWS (Amazon Web Services) for 5 years, cofounded Lilac.
The tech industry needs scalable infrastructure to support development of AI models, and Lucas envisions a completely new foundational layer between data scientists and their infrastructure. While Kubernetes is the current solution for high reliability for deployments, it is needlessly complex.
Lilac’s current release focuses on the GPU connection layer and is fully open source. For companies with on-prem clusters who need cloud failover, the platform addresses compute pooling across locations. Similar to Apache Airflow, Lilac’s business model is to build adoption via the free version and then monetize through an expert hosted cloud version.
Lucas’s long term vision for Lilac includes rewriting Docker runtime in Rust for better speed, security, and memory safety. This would enable cost-efficient, optimal model deployment for larger companies who are facing issues with instant scaling due to slow spin-up/down.
Getting in Touch with Lucas and Lilac
Right now, Lilac remains a 2-person team working an 8am to midnight, 7 days a week schedule during the YC 2025 Summer batch in San Francisco.
Lilac is looking for more production-scale testing with companies that require many GPUs across locations.
Reach out to Lucas via email at [email protected] or LinkedIn.