Next Gen AI Hackathon
Fri, Sept 22 - Sat, Sept 23 // San Francisco
What are the new mobile-first behaviors and use cases that generative AI will unlock?
We’re bringing together the most ambitious developers, engineers, hackers, and designers for a 48-hour hackathon to imagine mobile-first use cases that will create healthier, more productive, and more creative futures.
Participants will have the chance to share their concepts with leading AI investors, including Pear, Lightspeed, and Samsung Next; executives from top cloud providers; and select companies from our network of ecosystem partners.
There will be three award categories.
Breakthrough Award: The project that demonstrates outstanding deep core technical or architectural innovations that enable or leverage generative AI for mobile use cases.
Edge Award: The project with greatest impact that leverages only models that have been optimized and deployed locally on the device.
Pioneer Award: The project that most uniquely leverages mobile sensor, data, or services to create novel or more impactful mobile user experiences.
Each winning team will receive:
More than $100K in cloud credits from leading providers
A $1000 gift card to Samsung.com for each team member
The possibilities are endless but hackathon space is limited.
Applications are now closed.
OUR PARTNERS
The Details
Is there a specific challenge or sector we’ll be focused on?
The only requirement is that teams produce working demos that leverage generative AI to impact mobile device experiences. Beyond that, this is a blue sky challenge.
To get the gears turning, here are some thought starters:
How can generative AI augment on-device sensors like GPS, cameras, and microphones, etc. to deliver enhanced capabilities for applications?
How can generative AI improve how consumers interact with photos, emails, text messages, health data and deliver novel engaging experiences?
How can generative AI better extract user context and intent to improve interfaces like keyboards, voice, and app/menu navigations?
How can generative AI models be optimized and deployed on-device to enable or improve applications that are sensitive to privacy, latency, or network instability?
If you need even more ideas, we asked our friends and families about features they’re looking forward to seeing with the advancement of AI, and here are some that we thought were fun:
Health and wellness: Wellness/journaling app that correlates unstructured data from free form journal entries with wearable/health data
Creativity: Photo app that generatively edits photos based on user preferences (favorite selfies, photographic compositions, stylization)
Lifestyle: Generative bedtime picture book creator where your child is the main character
Generative interfaces: Android keyboard with generative functionality that adapts to different contexts: productivity vs. social apps, texts to your family vs. specific friends, intent-based adaptive menu options, etc.
What will we be expected to show by the end of the hackathon?
Participants are expected share a 3-minute pitch deck including a demo and high-level architectural diagrams. We will have mentors available to help folks fine tune and prepare for the final presentations.
Who should enter? What is the selection criteria?
We’re looking for individuals or teams (up to 4 members) from a diverse range of backgrounds who are passionate about AI and can leverage their experiences as hackers, engineers, designers, and storytellers to identify and prototype new use cases, interfaces, or behaviors that couldn’t exist without generative AI. We’re prioritizing submissions with a large vision and the potential to engage a broad set of mobile users on a frequent basis– either directly via novel application experiences, step function improvements to current user behavior, or indirectly via enabling technology.
All participants must be 18 years and older.
What is the application timeline?
Applications are now closed.
What is the judging criteria?
All projects will be evaluated on the following criteria.
Impact: Does this project have the potential to impact millions - even billions - around the world?
Creativity: Does the project solve a problem in a differentiated or unique way?
User Experience: Does the project deliver a delightful user experience?
Centrality of Generative AI: Does generative AI play a critical role in enabling the solution?
Technical feasibility: Does the technical implementation work? If the project is not fully functional, could this approach work technically?
Fidelity: How “finished” is the project? How close is it to an alpha launch?
Who are the judges?
James Currier, Founding Partner, NFX
Andy Duong, Investor, Samsung Next Ventures
Jeremiah Owyang, Investor, Blitzscaling Ventures
Lisa Han, Partner, Lightspeed Ventures
Aparna Sinha, Partner, Pear VC
Arthur Soroken, Head of Growth & Community, AI @ Google
Tool Chain Partners
Participants are encouraged to setup their environments and toolchains prior to the hackathon. To maximize your productivity during the event, we’ve also partnered with open source projects and toolchains that could accelerate your project development. We’ll be holding onboarding sessions and office hours where you can learn about each tool’s capabilities and ask the team questions. Join our Discord community for details.
BentoML + Open LLM // OpenLLM, by BentoML, is a framework for fine-tuning, serving, and operating production-ready Large-Language models (LLMs) via its integrations with LangChain, LlamaIndex, and BentoML.
Gradient // Gradient is a web-API based LLM dev platform that helps companies build private AI applications that they own by simplifying fine-tuning and inference via a single API call.
ChainML + Council // Council, by ChainML, is a robust AI agent platform enabling users to create sophisticated agents for complex tasks with seamless deployment and scalable oversight features.
MLC-LLM // MLC-LLM is a universal framework for developing, optimizing and deploying large models natively on-device, accelerated with local phone/laptop GPUs with no need for server support.
Qdrant // Qdrant is a high-performance, scalable, secure, open-source, similarity (vector) search engine and database, essential for building the next generation of AI/ML applications.
Tool Chain Virtual Onboarding & Office Hours
We’re hosting onboarding and office hours with our open source project and toolchain partners in the weeks leading up to the event. RSVP below if you're interested; a great resource for setting up the environments and toolchains for our mobile-first gen AI hackathon and beyond.
Aug 29 @ 3PT: Microsoft for Startups on how to provision and manage Azure Cognitive Services
Aug 31 @ 3PT: Qdrant on how to harness a high-performance, scalable, secure, open-source, similarity (vector) search engine and database
Sept 5 @ 3PT: MLC-LLM on how to develop, optimize and deploy large models natively on-device
Sept 7 @ 3PT: BentoML on fine-tuning, serving, and operating production-ready LLMs
Sept 12 @ 3PT: OpenVINO on optimizing and deploying AI inference on a range of Intel platforms
Sept 13 @ 3PT: Google on PaLM 2, a next-gen LLM that builds on Google’s legacy of breakthrough research in machine learning and responsible AI
Sept 14 @ 3PT: ChainML on the Council AI agent framework to empower devs to create and deploy customized, application-ready AI agents with predictable behavior
Sept 18 @ 3PT: AWS on all things building generative AI apps w/ AWS
Sept 19 @ 3PT: Gradient on how to personalize and build on open-source LLMs by simplifying fine-tuning and inference via a single API call
