What if one Saturday of building an AI project could end with you taking home an NVIDIA DGX Spark AI supercomputer worth $5,000?
That’s the main prize at the Zero Downtime Hackathon, an in-person AI event happening in San Francisco on August 22, 2026.
The challenge is built around a problem that becomes increasingly important once AI projects move beyond demos: reliability. Participants have to build an application, automated pipeline, or agentic workflow that can organize its architecture, work with changing web data, and expose what is happening when something fails.
There’s another reason this event may catch developers’ attention: projects will be judged by engineers from Anthropic.
The event allows participants to form teams beforehand or assemble one on-site. All building, demos, and submissions take place during the event.
Zero Downtime Hackathon at a Glance
| Feature | Information |
|---|---|
| Date & Time | August 22, 2026. 9:00AM-6:00 PM PDT |
| Location | San Francisco, California |
| Format | In-person |
| Team Size | 1-4 members |
| Grand Prize | NVIDIA DGX Sparx worth $5,000 |
| Judges | Engineers from Anthropic |
The Challenge Is More Than Building an AI Demo
The goal isn’t simply to create another chatbot.
Participants need to connect Port, Bright Data Scraper Studio, and SigNoz into a working system.
Port acts as the organizational layer, where teams define project goals, technical decisions, risks and tasks. Bright Data Scraper Studio handles web-data extraction, including demonstrating automatic scraper repair when a website changes its HTML structure.
SigNoz provides the observability layer, tracking things such as traces, metrics, logs, latency and failures.
Participants can build using coding assistants including Claude Code, Cursor and Codex. The result should demonstrate a complete lifecycle rather than a feature that works only in isolation.
What Does the Winning Project Need to Show?
The NVIDIA DGX Spark worth $5,000 goes to the project judged as the best full-stack integration.
The organizers specifically want to see Port, Bright Data Scraper Studio and SigNoz working together as one complete pipeline. Suggested project directions include a self-healing market intelligence system, an observed Docs-to-RAG bot and a self-healing CI/CD pipeline.
That gives participants a useful clue about what matters.
A flashy AI interface alone may not be enough. A stronger project should demonstrate how it is planned, how data enters the system, what happens when something breaks, how the system recovers and how developers can see what is happening behind the scenes.
There are also separate prizes for the best individual integrations with Port, Bright Data Scraper Studio and SigNoz. Each member of a winning track team receives a Keychron keyboard valued at $120.
Before You Register, There’s One Important Catch
This isn’t an online hackathon.
Participation is in person in San Francisco, so developers outside the area need to take the location into account before registering.
Teams can contain one to four people, and participants don’t necessarily need to arrive with a team already formed.
By the end of the event, each team must submit a GitHub repository containing its code, commit history and README. A three-to-five-minute demo video is also required, showing the terminal workflow, Port dashboard, live SigNoz monitoring and Bright Data auto-fix feature.
For developers in or around San Francisco, the Zero Downtime Hackathon combines several things that make an AI event worth watching: a practical production-focused challenge, modern coding assistants, judging by Anthropic engineers and a $5,000 NVIDIA DGX Spark as the grand prize.
Registration is currently available on the official event page. Check it before attending for the latest registration status or event changes.
