Inside the First Stanford Accel Leadership Program Hackathon with Anthropic and Perplexity

We recently hosted our first Stanford Accel Leadership Program (ALP) Hackathon, in partnership with Anthropic and Perplexity.

ALP is our flagship program for Stanford engineering students and aspiring entrepreneurs. Many of the most important entrepreneurial skills—solving problems, leading teams, making trade-offs under pressure—rarely show up in standard curriculum. ALP fills the gap for students with real-world reps in these areas.

The Hackathon’s Goal: Hands-On Practice

For the inaugural Hackathon, we brought together ALP students and alumni who have founded startups and joined companies like Waymo, Loom, and Linktree. The goal was straightforward: offer a chance to move from idea to working product under tight constraints.

With the barrier to shipping product lower than ever, we see speed of execution, judgment about what to build, and the ability to rally a team as the skills setting apart today’s strongest founders. This was an opportunity to put them to use.

To kick off the day, Accel portfolio companies Anthropic and Perplexity walked through best practices for their tools, then equipped participants with Claude Code credits and exclusive beta access to Perplexity’s newest Search API.

‍Perplexity’s Kesku Saif unveiled the latest Search API, allowing users to choose any model to build into their own LLM. — ‍*Perplexity’s Kesku Saif unveiled the latest Search API, allowing users to choose any model to build into their own LLM.*

‍

*Anthropic’s Thariq Shihipar and Kashyap Murali conducted a live demo of a 5-minute quick start on Claude Code.*

‍

*Seven teams took over the Accel Palo Alto office for the day, stopping only briefly for lunch and dinner.*

‍

*The pitches drew big reactions from our audience and judges.*

‍

A Startup Journey, Compressed into a Day

The day’s theme was Hack Your Own Headache, and teams picked problems they’d encountered in their own lives. Projects were wide-ranging, including a voice-enabled banana thief stopper, AI audio fraud detection, and a multi-harness agent research lab. Several teams riffed independently on a similar idea of storing our digital footprint in local memory.

For participants, the Hackathon day followed the typical course of a startup’s journey: the initial excitement, the waning novelty, the trough of nerves, the hint of promise—and the caffeine-fueled sprint to the deadline, featuring inspired coding and frenzied pitch practice. The late burst of energy felt familiar for anyone who’s experienced a launch firsthand.

The Inaugural Stanford ALP Hackathon’s Winning Projects

The judges scored projects on a six-part rubric, and the top five teams finished within just a few points of each other.

In the end, three projects edged out the rest:

First Place: Voicera (AI fraud detection)
Second Place: Daily Brief (custom daily podcasts)
Third Place: Trail (local, personal memory)

*Hats off to Voicera for their winning project.*

‍

Thank you to everyone who joined us for this first-of-its-kind event. And a special thank you to our speakers and judges Thariq Shihipar and Kashyap Murali, Anthropic technical staff, and Kesku Saif, Perplexity’s head of community.

‍