Limited Edition AI Hackathon 2026 Japanese Tumblers

Prizes:
1st place: $1500 Cash Prize + Internship Interview Opportunity
2nd place: $1000 Cash Prize + Internship Interview Opportunity
3rd place: $500 Cash Prize + Internship Interview Opportunity

Most AI applications stop at conversation. Your challenge is to build an AI agent that can be discovered through ASI:One, understand a user’s intent, and take meaningful action to solve a real-world problem. Your agent might coordinate services, automate a workflow, analyze live information, make recommendations, complete transactions, or collaborate with other specialized agents. The problem and approach are up to you, but the result should be more than a chatbot or a thin wrapper around an API.

Prize: $5000 in API credits (total), an office hour with our Applied AI team, and an invitation to visit our office in SF.

Our tech prize this year celebrates teams that reach the furthest: projects built with Claude Code that tackle meaningful issues in health, education, economic opportunity, or any domain where AI could genuinely shift what's possible for people. We know you have limited time, but aspiration and effort matter more than the outcome this time, so take the biggest swing toward the most challenging problem you can think of!

Prize: Mac Minis, 25k Redis Cloud credits for the team, Redis Timbuk2 backpacks

- Using Redis Beyond Caching: leveraging our AI tools like Redis Iris for Agent memory, vector search, context retrieval
- Creativity and originality: solving real human problems in ways we haven't seen before, has a fun factor
- Technical Implementation: Quality and sophistication of the engineering, Correctness, scalability, and architecture.

Prize:
1st) $2000, Claude Code 5x 6months, Interview
2nd) $1000

Depth of research, ingenuity, creativity.

1st Place: $2000 (split amongst team)
2nd Place: Merch + Credits
3rd Place: Merch + Credits

Participants are tasked with building any agent that uses the web - any framework, harness, model - but it must be powered by the Browserbase platform. Must use one of the following: browsers, search, fetch, Stagehand, Browse CLI. Prize TBA.

1st place - $750 + campus tour + interview opportunity
2nd place - $250 + campus tour + interview opportunity

Goal: Make the agent reliable in production environments

Criteria:
1. How did they use Agentspan to build an agent or bring an agent in (LangGraph, Google ADK, Open AI SDK, or any other framework)
2. How well did they integrate Agentspan into their demos
3. Explain what makes it reliable in production

Prize: Prize package worth $500

Projects must use a Simular product (Sai, SimuLang, or Agent S) in a meaningful way and post about the experience on X or LinkedIn, tagging @official accounts. Judges will pick the winning team based on creativity (bonus points for novel use cases), technical execution, and real-world impact. (Optional if the organizer has other submission channel - to be considered, email zening@simular.ai with your team name, project description, social post link, and a short note on how you used Simular product).

Prize: $1000 main winner, 2 x $250 runner-up prizes

Hard requirements:
- Product uses QNX OS

Judging criteria:
- Is it a ”cannot-fail” embedded application?
- Does it require real-time or reliability?
- Is it using AI in an interesting or creative way?
- Is it running on the embedded hardware (and not in cloud)?

Prize: $1,000 cash (delivered as gift cards)

Evidence that Arize was used and actually improved the application.

Prize: 4 x Nintendo Switch 2 for each member of the team

The Deepgram prize goes to the team that builds the most creative and well-executed voice-powered experience using Deepgram's APIs. We'll judge on creativity of the voice interaction, technical execution, and how essential voice is to the experience (rather than tacked on). Submitted projects must demonstrably use at least one Deepgram product (speech-to-text, text-to-speech, or the voice agent API).

Prize: 4x Nintendo Switch 2 System for each member of the team and Guaranteed Interview

Sentry is looking for builders who go beyond the prompt, we want to see strong technical execution paired with clear communication, collaborative problem-solving, and the confidence to speak up, course-correct, and lead when it counts. The best submissions will demonstrate not just what you built, but how your team worked together under pressure to make it happen. Bonus points if you leveraged observability or error monitoring in your project(s), we love engineers who think about reliability from day one.

Prizes:
First: $500 + office visit + Pika MCP credit 10k + special merch;
Second: $300 + office visit + Pika MCP credit 10k
Third: $200 + office visit + Pika MCP credit 5k

Pika is the AI company making creation possible—and fun—for normal people. Whether that’s through prompt-based video generation, unreal effects on your existing content, or a personal agent who you collab with, we’re always cooking up the next way to create. We’ll be hosting a workshop on how to turn Claude into a creative partner through our MCP, and offering cash rewards, product credits, office tours, and company swag to people whose hackathon ideas make us the most jealous.

We value:
•⁠ ⁠Creativity and originality
•⁠ ⁠Effective use of Pika MCP and AI workflows
•⁠ ⁠Quality and impact of the final output

Prize: Sunset Cruise (+30 friends invited) + $500 Devin Credits
Most interesting & technically impressive project built with Devin.

Prize: $1000 overall in amazon gift cards, split equally to winners. Prize to be distributed post hackathon.
We would like to provide the prize to the top project overall that used Band as a key technology in their project. So we rely on the regular judging for ranking projects, just pick the top one that used Band in a significant way (at least 2 agents collaborating via the BAND platform).

1st place - $1500
2nd place - $1000
3rd place - $500

The top 4 Ghost Trials performers invited to perform on stage at our June 26 SF show!

Criteria:
Open to any team, stack, or direction, as long as it's physical AI: code that closes the loop with a real or simulated robot, or generates, curates, or evaluates the data and worlds that train one. Submissions via Devpost on Sunday (≤3-min demo video, public repo or doc, 200-word writeup); the judges' test: would a real robotics team use it?

Prize: Tiny tapeout credits
Projects will be evaluated on their creative and technically sound use of the Cognichip Artificial Chip Intelligence™ platform to solve a real chip design problem. Judges will assess design methodology and execution, quality of the final deliverable, and how effectively the team leveraged AI to accelerate the design process.

Prize: 1000$ (1st place), 400$ (2nd place)
Build a simple annotation app, use Terac to collect real human-labeled data from our panel, then fine-tune a model on that data and show it beats the base model. To be eligible, your training data must come from annotations collected through Terac during the event, not synthetic or pre-existing datasets. Projects are judged on improvement over the base model shown to hold on unseen examples (50%), the creativity and UX of your annotation environment (30%), and how smartly you used the human data within your $250 credit (20%).

1st place - $1000 credits
2nd place - $500 credits
3rd place - $300 credits

Guaranteed admission to SkyDeck’s Pad-13 incubator program

Kodak Chemera Cameras

XGIMI Projector

Flashforge 3D Printers

$100 Cal Student Store Gift Cards

AirPods 4 Noise Cancellation

GMMK Keyboards

1st Place - Nintendo Switch