Welcome to the AI 4 Infra Challenge — Beginner Track  DARK DATA CHALLENGE

This track is built for anyone curious about AI and data — no prior coding or machine learning experience required. You'll work with real municipal work-order records and crew notes to explore how AI and data analysis can turn "dark data" into useful maintenance insights. Organizers and mentors are on hand throughout to help you get unstuck, and a starter notebook is provided so you're never starting from a blank page.

This challenge, you'll analyze work orders from: Water and sewer, streets and traffic, electric systems, municipal buildings, reconnect related jobs, identify recurring problems, and recommend preventive actions

Event Schedule

  • Kickoff — Thu, Sept 3, 4:00–6:00 PM, Student Commons 1600
  • Submissions open — through Sept 11, 11:00 AM
  • Demo Day / Judging — Fri, Sept 11, 11:00 AM–3:30 PM, North Classroom
  • Awards Ceremony — Fri, Sept 11, 4:00–5:00 PM, North Classroom 1005
  • Industry Symposium (winners showcase) — Wed, Sept 16, 9:00 AM–4:00 PM, Jake Jabs Center

Hosted by the AI Student Association at the University of Colorado Denver, in partnership with Colorado State University, Trimble, WSB, HDR, Colorado Smart Cities Alliance, and SHPE.

Requirements

THE CORE REASONING TASK

You must infer:

  • Which jobs describe the same problem
  • Whether an issue is recurring
  • Whether the problem was resolved
  • What may have caused it
  • What preventive action should follow

Each team should submit a brief presentation or written summary (slides, doc, or notebook) covering:

  1. Problem and scope — the maintenance question, asset group, or workflow component you chose, and why it matters
  2. Proposed approach — the records, tools, analysis methods, and intended outputs you used
  3. Evidence of exploration — relevant work-order examples, observations, diagrams, experiments, code, or prototype work
  4. Example insight or intended output — a finding, sample report, or clearly labeled mockup. Reference real work-order IDs where you use them, and clearly separate observations from hypotheses
  5. Documented workflow — what you completed, how you approached it, and what remains
  6. Limitations and next steps — what you learned, what's still uncertain, and how you'd continue

You do not need a complete pipeline, a calibrated confidence score, or an end-to-end report — a well-reasoned partial submission scores just as well as a "finished" one. On Devpost, submit your write-up/slides plus any supporting code, notebooks, or screenshots.

Hackathon Sponsors

Prizes

$1,000 in prizes
1st Place
$500 in cash
1 winner

2nd Place
$300 in cash
1 winner

3rd Place
$200 in cash
1 winner

Devpost Achievements

Submitting to this hackathon could earn you:

Judges

Farnoush

Farnoush
CU Denver

Judging Criteria

  • Rubric
    Will be sent over email! Email aisa@ucdenver.edu for more info if you haven't recieved it!

Questions? Email the hackathon manager

Invite others to compete

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.