merge(happy_hour) → self_improving_agents @ SGLang Summit

Celebrate SGLang Summit and AI Engineer Code Summit Together

Your agents improve every run. Your night improves every round.

Most agents ship once and stay the same. The best ones keep getting better. They learn from every run, every eval, and every piece of user feedback.

Join DigitalOcean, NVIDIA, SGLang, and Major League Hacking for an evening about building that loop. You’ll see how teams build agent harnesses, safely run coding agents in sandboxes, measure them against their own traffic instead of a public leaderboard, route each task to the right open model, and turn what they learn in production into the next, better version of their agent. It all happens on one platform, from inference to agents.

Expect lightning talks and live demos, food, drinks, and conversation with engineers, researchers, and founders who are running agents in production.

Who should come: Anyone building agents, including:

• AI and ML engineers
• Infra engineers
• Founders
• Technical leaders who want their agents to get better after launch

Agenda

• 6:00 PM: Doors open
• 6:30 PM: Lightning talks and demos
• 7:15 PM: Food, drinks, and fun

Speakers

• Jiantao Jiao, Director of AI & Distinguished Scientist, NVIDIA
• Balaji Varadarajan, Inference Engineer, DigitalOcean
• More speakers to be announced; stay tuned!

Space is limited and registration is by approval. Request your spot.

Hosted by DigitalOcean, NVIDIA, SGLang, and Major League Hacking

Note

Alcohol will be served at this event. Attendees must be 21 or older and will be asked to present valid government-issued photo ID at check-in. Non-alcoholic options will be available.

DigitalOcean is the organizer of this event and will use the information you provide to register you, communicate with you about the event, and manage on-site logistics. See the [DigitalOcean Privacy Policy](https://www.digitalocean.com/legal/privacy-policy) for how we handle your information and your privacy rights.