Build with AI Hackathon 2027

Build something useful. Prove that it works. Make it trustworthy.

Bring a real problem, a curious mind, and a willingness to experiment. Leave with a working prototype, evidence of what you learned, and a project you can explain with confidence.

Build with AI is a three-day online hackathon where data, intelligent applications, and cybersecurity come together. Move beyond an impressive prompt or a polished slide deck: build a solution that helps someone, demonstrate its value, and show how you have considered reliability, privacy, and safety.

Event at a Glance

Register for this event to receive confirmed event updates. Every participant registers individually, including team members. Registration for another DevBuild Studio event does not register you for this hackathon.

The Challenge: From Idea to Evidence

Choose a problem worth solving, identify who experiences it, and build the smallest solution that demonstrates a meaningful improvement.

A strong entry answers five questions:

1. Who is this for? Name a specific user and the problem they face.
2. Why this approach? Explain why your data, AI, or security solution fits the problem.
3. What actually works? Demonstrate a functioning end-to-end workflow.
4. How do you know? Show a baseline, test result, or other concrete evidence.
5. What could go wrong? Explain limitations, potential misuse, and safeguards.

An ambitious idea is welcome. A smaller, well-tested prototype is better than a broad promise that cannot be demonstrated.

Three Themes, One Shared Goal

Choose one primary theme so judges can understand your project's focus. Projects may connect multiple themes; explain those connections in your submission.

Theme 1: Data - Turn Information into Better Decisions

Build something that makes data more useful, reliable, understandable, or accessible.

Challenge directions

• Turn an openly licensed dataset into an interactive tool that helps a specific audience make a decision.
• Build a data-quality assistant that detects missing values, inconsistent records, or suspicious changes and explains its findings.
• Create a reproducible analytics pipeline with clear provenance, validation, and a useful output.
• Make a complex dataset accessible through plain-language explanations, inclusive visualizations, or an assistive interface.

Show your evidence: Explain where the data came from, whether you may use it, how you checked its quality, and what changed compared with a simple baseline. Do not confuse correlation with causation.

Example: A community-resource navigator built from public data, with freshness checks and visible source links instead of unsupported recommendations.

Theme 2: AI - Build Intelligence People Can Trust

Create an AI-powered experience that solves a concrete problem and makes its capabilities and boundaries understandable.

Challenge directions

• Build a grounded assistant that cites approved sources and can say when it does not have enough information.
• Design a human-in-the-loop workflow that helps a person complete a task while keeping consequential decisions under human control.
• Create an evaluation tool that measures quality, failure cases, latency, or cost rather than relying on a handful of impressive outputs.
• Build an accessible learning or productivity tool that gives people useful control over AI-generated suggestions.

Show your evidence: Explain why AI is needed, compare against a non-AI or simpler baseline, test representative failure cases, and describe how you handle uncertainty, harmful output, and data exposure.

Example: A document assistant that links every answer to a source, refuses unsupported claims, and includes a small, reproducible evaluation set.

Theme 3: Cybersecurity - Make Digital Systems Safer

Build defensive tools, safer development workflows, or understandable security experiences. All security work must stay within explicitly authorized environments.

Challenge directions

• Analyze synthetic security logs to help a human identify and investigate suspicious patterns.
• Create a secure-by-design assistant with input validation, permission boundaries, and defenses against untrusted instructions.
• Build a developer tool that identifies exposed secrets or risky dependencies in a repository you own or have permission to assess.
• Make security guidance easier to act on through a clear, accessible interface that prioritizes remediation.

Show your evidence: Define the assets and threats, state your testing authorization, describe false positives and blind spots, and demonstrate defensive value without exposing secrets or attacking third-party systems.

Example: A local security-review companion that examines a deliberately prepared sample repository and explains how to resolve findings.

Who Should Participate

Students, developers, data practitioners, designers, cybersecurity enthusiasts, researchers, and curious builders can contribute.

You do not need to be an expert in all three themes. Bring one useful perspective: understanding a user, exploring a dataset, writing code, designing an accessible interaction, testing an assumption, or explaining a result.

First hackathon? Start with a small problem and a demonstrable workflow. Experienced builder? Push the technical depth while keeping the project focused and reproducible.

Participation remains subject to applicable law, any required guardian consent, and the final event eligibility notice. University affiliation is not a substitute for individual registration.

What You Can Build With

Use the languages, frameworks, models, and tools that best suit your idea. No particular AI vendor or technology stack is required.

• Use open-source tools and openly licensed or otherwise authorized datasets.
• Use local models, commercial APIs, traditional analytics, or a combination.
• Include a simple setup path so someone else can understand and review your work.
• Keep paid services optional where practical. Disclose any accounts, credentials, hardware, or spending required to reproduce the demo.

Free event registration does not include a promise of cloud credits, paid software, hardware, or reimbursement. Use spending limits and do not expose API keys.

Team Formation and Collaboration

You may participate individually or collaborate with a team. Small, complementary teams are encouraged.

• Every contributor must register for this specific event.
• List all team members and describe each person's contribution.
• Each participant belongs to one competing team; each team submits one primary project.
• Choose a primary theme. Explain additional theme connections without submitting duplicate copies of the same project.
• Credit external advice, reused components, and AI assistance.
• Do not claim another person's work or submit the same project under different team identities.

Any formal team-size limit or additional eligibility restriction will be announced before team finalization, not introduced during judging.

Three-Day Program

The following is the program outline, not a timed agenda. Exact session times, required attendance, the build window, and submission deadlines will be published in Pacific Time before kickoff.

Friday, February 19 - Understand, Choose, and Start

• Review the event brief, participation conditions, and submission checklist.
• Choose a primary theme and define your user, problem, and success measure.
• Confirm your team and document any pre-existing components.
• Establish a simple baseline, plan your prototype, and begin work when the official build window opens.

Target outcome: A clear problem statement, a focused scope, and a first working path.

Saturday, February 20 - Build, Test, and Improve

• Connect the core workflow from input to useful output.
• Test with authorized, synthetic, or appropriately licensed data.
• Evaluate representative cases, including failures and edge cases.
• Improve usability, accessibility, privacy, and security.
• Record results and prepare a repeatable demo.

Target outcome: A functioning prototype with evidence, not just a concept.

Sunday, February 21 - Demonstrate, Submit, and Reflect

• Complete the prototype within the announced build window.
• Prepare the project page, demo, setup instructions, and disclosures.
• Submit through the official project-submission channel before its published deadline.
• Be available for any announced judging questions or demo session.
• Explain what you achieved, what did not work, and what you would do next.

Target outcome: A complete, reviewable submission that tells an honest story.

Take breaks and work sustainably. Three event dates do not mean that participants are expected to stay online continuously.

Prepare Before the Event

• Register individually and check that your contact details are correct.
• Arrange a computer, reliable internet access, and a way to demonstrate your project.
• Set up your development environment, version control, and any approved accounts.
• Review dataset and dependency licenses.
• Explore problem areas and form a team, while respecting the official build-window rules.
• Check API quotas, spending limits, and privacy settings.
• Plan an accessible demo: readable text, understandable visuals, and captions or a transcript for recorded narration.

Meeting links, the submission channel, session timings, and any supported resources will be announced separately. Mentor availability, workshop sessions, and recordings are not yet confirmed.

Project Submission Checklist

Submit one project entry per team. A registration confirmation is not a project submission.

Your entry should include:

1. Project name and primary theme.
2. Team and contributions: registered participants and what each person built.
3. Problem and user: who needs the solution and why.
4. Working demo: a concise walkthrough of the actual prototype, with an accessible recording or equivalent explanation where practical.
5. Code or implementation access: a repository or reviewable package, plus instructions. If private, provide authorized judge access through the announced channel.
6. README: architecture, dependencies, setup, and a reproducible example.
7. Evidence: baseline, evaluation method, results, and known limitations.
8. AI disclosure: models and tools used, their role, and which work was generated or assisted.
9. Data and license disclosure: sources, permissions, attribution, and pre-existing components.
10. Responsible-build note: privacy, security, accessibility, potential misuse, and safeguards.

Do not include passwords, API keys, private customer information, student records, or credentials in your submission. Remove secrets from both current files and repository history before sharing.

The official submission form, demo-length guidance, and exact cutoff will be announced before the event. Do not send a project to an unrelated event page or assume that uploading a file elsewhere completes submission.

Judging: A 100-Point Rubric

The same rubric applies across all three themes. Judges assess the work demonstrated, not the number of technologies listed or the polish of unsupported claims.

Projects must meet eligibility, conduct, and safety requirements before scores are considered. An unsafe or unauthorized project cannot compensate for a rule violation with a high technical score.

For equal totals, compare validation and evidence first, then technical execution, then practical impact. Any further tie requires a documented panel decision using the published criteria.

Judges should disclose material conflicts of interest and recuse themselves where needed. Organizers and judges must not evaluate their own work. Clarifications must not become an opportunity to add new development after the submission deadline.

The judge roster, evaluation format, and results timing will be announced. No individual judge, sponsor, or institutional endorsement is implied by an unannounced role.

Prizes, Recognition, and Project Showcase

Build for the value of the problem and the quality of your work.

No cash prizes, sponsor awards, certificates, internship opportunities, or other benefits have been announced. If awards are introduced, the organizers will publish their value, eligibility, selection process, and any additional conditions before they apply.

A public project showcase may be arranged separately. Sharing your project publicly or using your name, image, or demo for promotion requires the applicable permissions; it is not automatic permission created by registration.

Participation Conditions

Original Work and the Build Window

Competition development must take place within the officially announced build window. Before it opens, you may explore ideas, form a team, and prepare general-purpose tools.

Existing libraries, frameworks, models, and permitted datasets may be used with attribution. Clearly identify pre-existing project-specific code or assets. Only the team's new work within the build window is eligible for evaluation.

Do not present an existing product, a lightly relabeled application, purchased work, or another person's project as a new hackathon build. Stop competition development at the announced submission cutoff.

AI-Assisted Development

AI tools are welcome when used responsibly and transparently. Explain what they did, what your team contributed, and how you checked the result.

You remain responsible for generated code, data handling, dependencies, factual claims, and compliance with licenses and service terms. AI output is not proof of correctness or a replacement for testing.

Do not submit private prompts, confidential source material, credentials, or personal information merely to demonstrate AI usage.

Data, Privacy, and Security

Use synthetic, public, or explicitly authorized data. Document provenance and restrictions. Do not upload confidential, regulated, or personally identifying information without a lawful basis and the necessary permissions.

Security testing is limited to systems you own or have explicit written authorization to assess. No unauthorized scanning, phishing, credential collection, exploitation, or disruption of third-party systems is permitted.

Do not use the event to distribute malware, expose private information, or encourage harm. Keep demonstrations bounded, defensive, and safe for reviewers.

Intellectual Property and Third-Party Rights

Participation does not itself transfer ownership of your project to the organizers. Ownership remains subject to contributor agreements, employer or university obligations, and third-party rights.

Submit only material you are entitled to share. Identify licenses and respect attribution, model-use restrictions, dataset terms, trademarks, and other applicable rights.

Submitting a project permits organizers and assigned judges to access the submitted materials for event administration and evaluation. It does not automatically grant commercial rights or blanket promotional permission. Open-source publication is optional unless a separately announced challenge expressly requires it.

Fair Participation and Enforcement

No plagiarism, fabricated results, deceptive demos, duplicate team entries, impersonation, or attempts to manipulate judging are allowed. Disclose material assistance and any overlapping use of the project in another competition.

Organizers may request reasonable evidence to verify a submission and may exclude entries that violate published rules. Where practicable, participants should receive an explanation and an opportunity to clarify facts. Immediate restrictions may be necessary to address a safety concern.

Report administrative mistakes or procedural concerns through the official event contact channel. Do not pressure judges or other teams.

Costs, Changes, and Withdrawal

Registration is free. Participants are responsible for their own connectivity, equipment, and any optional services they choose to use unless a benefit is explicitly announced.

Material changes to dates, format, eligibility, deadlines, or award conditions should be communicated through the event page and registered-participant updates. Competition rules should not be changed retroactively to disadvantage completed work.

If you cannot attend, notify the organizers. These participation conditions do not remove any rights that cannot lawfully be excluded.

Code of Conduct

Help make the event a place where people can build without being excluded, intimidated, or harassed.

• Treat participants, organizers, reviewers, and guests with respect.
• Welcome different backgrounds, identities, abilities, and levels of experience.
• Give constructive feedback about the work, not personal attacks.
• Do not harass, threaten, discriminate, stalk, share private information, or record people without the required consent.
• Respect requests to stop an unwelcome interaction.
• Do not retaliate against someone who reports a concern.
• Follow moderation instructions across event meetings, messages, shared documents, and project demonstrations.

To raise a concern, contact the organizers through the DevBuild Studio contact page. Share sensitive details privately, not in public project comments. An event-specific reporting contact will be announced before kickoff.

Organizers may intervene, restrict access, or remove a participant to protect safety. If there is immediate danger, contact the appropriate local emergency service.

Accessibility and Inclusive Participation

If you need an accommodation or an alternative way to participate, contact the organizers before the event where possible. Explain the support you need without posting medical details publicly.

For your own project and demo:

• Make core controls usable with a keyboard.
• Use clear labels, readable text, and sufficient contrast.
• Do not communicate essential information through color alone.
• Add meaningful text alternatives for important visuals.
• Provide captions or a transcript for prerecorded narration where practical.
• Explain any known accessibility limitations honestly.

Specific event accommodations, captioning arrangements, supported platforms, and recording availability must be confirmed with the organizers; they are not guaranteed by this announcement.

Frequently Asked Questions

Is registration free?

Yes. Optional third-party products or services may have their own costs; free registration does not cover those costs.

Is this event online or on campus?

Online. Physical attendance at a university campus is not required by this announcement. Joining instructions will be published separately.

Which time zone should I use?

Pacific Time: America/Los_Angeles. These February dates fall in PST, UTC-08:00. Exact session times and the submission cutoff will be stated in that time zone.

When does participant registration close?

Registration closes when the event's final UTC calendar day ends: February 22, 2027 at 00:00 UTC, which is February 21 at 4:00 PM PST, unless the 500-participant capacity is reached earlier. This registration cutoff is separate from the project-submission deadline, which will be announced.

Do I need to know AI, data engineering, and security?

No. Choose one primary theme and a problem you can address. Cross-theme projects are welcome, but mastering every area is not a prerequisite.

Can I participate without a team?

Yes. You can build individually or with a team. Every contributor must register individually. Any formal team-size restrictions will be announced before teams are finalized.

Can I use ChatGPT, coding assistants, or existing models?

Yes, subject to the applicable tool and model terms. Disclose their use, explain your contribution, test the output, and keep private information out of unapproved services.

Can I bring an existing project?

You may disclose and reuse permitted components, but you must not present an existing project as a new build. Judging considers new work completed within the announced build window.

Do I have to publish my source code openly?

Not under this announcement. Judges need sufficient authorized access to review the implementation. A private repository or equivalent reviewable package can be used through the announced submission process.

Is a slide deck enough?

No. Show a working prototype or a reviewable implementation, explain its limitations, and provide evidence. Slides may support a demo but cannot replace it.

Do I have to remain online throughout all three days?

No. Work sustainably. Exact required sessions and any judging availability will be announced before the event.

Are there prizes, certificates, mentors, or free cloud credits?

None have been confirmed. Check the event page for official updates rather than assuming that these benefits are included.

Does registering submit my project?

No. Individual registration and team project submission are separate steps. Follow the official submission instructions when they are published.

How do I ask a question or request an accommodation?

Use the DevBuild Studio contact page. Event-specific communication and reporting details will be announced before kickoff.

About the Organizers

DevBuild Studio and University of Washington Bothell are the event organizers. Official logistics, participation updates, and any approved supporting organizations will be identified through the event page and organizer communications.

Do not assume that a technology mentioned in a project is an event sponsor, or that a university affiliation grants access to private systems, datasets, or accounts.

Ready to Build?

Choose a problem that matters. Bring a small, testable idea. Build with care, demonstrate with evidence, and leave the project easier for someone else to understand.

Join Build with AI Hackathon, February 19-21, 2027. Register individually on this event page and watch for the confirmed schedule and submission instructions.

Guidance Behind These Rules

These event-specific conditions are written for Build with AI, informed by publicly available MLH guidance on hackathon rules, judging, submission planning, and community conduct.

These references do not imply MLH membership, sponsorship, or endorsement. Event-specific organizer instructions, not another organization's reporting contacts or contest terms, govern this event.