Hack theHarvestRegistration
Hosted and Organized by HMCI

Hack the Harvest

Can you tell what happened in a cornfield from space?

A two-day interdisciplinary hackathon exploring how satellite imagery and GPU-powered analysis can deliver faster, more accurate commodity crop projections than today’s survey-based process. Your starting challenge: distinguish silage corn from grain corn using publicly available geospatial data.

November 14–15, 2026 | Aggie Square, Sacramento, CA

200

Student Places

5–7

Students per Team

22.5h

Build Window

The ChallengeSilage or Grain?
Illustrative aerial view of adjacent green and golden agricultural fields
Silage Corn

Typically harvested earlier, while green

Grain Corn

Generally harvested later, after dry-down

In Collaboration With

USDA logo

Challenge Provider

UC Davis logo

Academic Host

NVIDIA logo

Technology Partners

Solidigm logo

Technology Partners

FarmGPU logo

Technology Partners

Keeboh logo

Powered By

The bigger goal: better crop projections.

Use satellite imagery and GPU-powered analysis to work toward faster, more accurate commodity crop projections than today’s survey-based process.

The Challenge

How can we tell from the sky whether a cornfield was harvested for silage or for grain?

Challenge Provided by USDA

Your task is to distinguish cornfields harvested for silage from those harvested for grain using only publicly available geospatial data. Identifying how a crop is harvested is a concrete step toward the broader crop-projection goal.

The core problem is detecting subtle differences in crop growth and harvest patterns from afar. The signal may be subtle, noisy, and easy to miss. But it is there.

Key insight

Silage corn is typically harvested earlier, while the crop is still green. Grain corn generally remains in the field longer as it dries down before harvest.

One shared USDA challenge. Different ways to solve it.

There is no single prescribed approach.

Approach 01

Changes Over Time

Look for changes in vegetation and compare satellite imagery across harvest windows.

Approach 02

Patterns in the Data

Combine public datasets, build field-level time-series models, or identify spectral and temporal signatures.

Approach 03

An Answer People Can Use

Develop a classification model, create an interpretable visualization, or find a completely different signal.

The challenge is yours to solve.

Who Should Hack?

You don't have to be a coder to hack the harvest.

We are looking for teams with different ways of thinking.

Engineers

Build the pipeline.

Work with sensors, systems, algorithms, computation, and technical implementation.

Data Scientists

Find the signal.

Analyze geospatial and time-series data, develop models, and test hypotheses.

Business & Management

Define the problem.

Develop strategy, identify opportunities, communicate insights, and think about practical value.

Open to undergraduate, graduate and doctoral UC Davis students across disciplines, including agriculture. Teams of 5–7 form before the event.

Bring one. Find teammates who bring the rest.

AIMachine LearningPythonData ScienceSatellite ImageryGeospatial AnalysisAgricultureBusiness StrategyProduct ThinkingVisualizationEntrepreneurshipSystems Engineering

What You'll Work With

Public data, local insight, and a sharp question.

Satellite Data

Earth-observation imagery and publicly available satellite datasets.

Crop & Agricultural Data

Public datasets describing crops, fields, and agricultural activity.

Geospatial Data

Maps, field boundaries, geographic information, and spatial patterns.

Time-Series Analysis

Look for changes in fields over time and identify harvest signatures.

AI & Machine Learning

Build models or analytical methods to distinguish silage from grain.

Creative Problem Solving

Find unconventional signals or combine multiple public datasets.

You do not need to be an expert in every one of these areas.

Why This Matters

A small difference. A big question.

Agriculture is happening all around us, but much of it can be difficult to observe at scale. Satellites and open data give us a new perspective: the ability to observe patterns across thousands of fields without being physically there.

Your job is to turn those observations into an answer.

Event Platforms

Where the event happens online

Devpost

Registration, team formation, project and video submissions, and judging.

Keeboh

Technical workspace and challenge resources.

Discord

Announcements, team coordination, and scheduled support.

The Hackathon Experience

Two days. One real challenge.

Draft schedule. All times Pacific.

Aggie Square event operations run 8 a.m.–4 p.m. both days. Participant check-in begins Saturday at 9 a.m.; teams are expected back Sunday by 10 a.m. The official build window is Saturday 10:30 a.m.–Sunday 9 a.m. (22.5 hours).

Pre-event orientation

Approximately one week before the event, a one-hour online orientation will cover the challenge, Keeboh, Devpost submissions and questions. Background materials, prior-work summaries, starter notebooks and access instructions will be shared in advance. Timing and session details will be announced once confirmed.

Prepare your team, read the background materials and test access in advance. No competition work may begin early. Work on competition solutions, challenge-specific model training and judged deliverables begins Saturday at 10:30 a.m.

Saturday, November 14 | Aggie Square

Discover & Build

4480 2nd Ave, Sacramento

9 a.m.

Check-in opens: refreshments and partner booths

10:30 a.m.

Your 22.5 hours of building begin

12:30 p.m.

Lunch with partners

3 p.m.

Overnight workspace opens at the Courtyard by Marriott Sacramento Midtown

Sunday, November 15 | Aggie Square

Build & Show

4480 2nd Ave, Sacramento

9 a.m.

Submissions close in Devpost

10 a.m.

Back at Aggie Square for sponsor sessions and career conversations

11:30 a.m.

Every team member present; finalists announced

12:15 p.m.

Live finals begin

1:50 p.m.

Awards and closing (awards may start as late as 2 p.m.; done by 2:30)

Between milestones, the time is yours to build, get mentoring and meet partners.

Overnight | Keep Building

The overnight workspace is at the Courtyard by Marriott Sacramento Midtown, 4422 Y St, Sacramento, CA 95817. It is open from Saturday at 3 p.m. through Sunday submissions at 9 a.m., followed by pack-up and return to Aggie Square from 9–10 a.m. Participants arrange their own transportation.

Mentoring and technical support will combine scheduled on-site and remote availability, with event updates and support coordination through Discord.

Sunday attendance requirement

Teams are expected back at Aggie Square by 10 a.m. To remain eligible for overall awards, every team member must check in individually and be present at Aggie Square by 11:30 a.m. Finalists must be ready at 12:15 p.m., and teams must remain through the awards ceremony. Emergency or accessibility exceptions require approval from the event lead.

Submissions & Judging

Show your work clearly.

Submit your project and a demo or pitch video of no more than three minutes through Devpost by Sunday at 9 a.m. Four eligible teams advance to the live finals, with seven minutes to present and three minutes for questions.

Final judging rules and prize details will be shared before the event.

Draft rubric

Technical Approach and Evidence30%
Relevance and Practical Value to USDA20%
Creativity and Innovation25%
Presentation and Demonstration25%

Award categories and prizes will be announced once confirmed. A separate NVIDIA Nemotron award is proposed, subject to confirmation. Nemotron use is optional for the overall competition.

FAQ

Before you arrive

Who can participate?

Undergraduate, graduate and doctoral students across disciplines, including engineering, computer science, data science, business, management and agriculture. Open to UC Davis students only.

Do I need to know how to code?

No. Teams benefit from a mix of technical, analytical, business and creative skills.

Do I need prior satellite-data experience?

No. Background materials, starter notebooks and a pre-event orientation will help participants prepare.

Do I need a team?

Teams of 5–7 form before the event. Team formation will be coordinated through Devpost. Team guidance will be shared with registration.

What data can we use?

Publicly available geospatial data. Challenge resources and technical instructions will be shared before the event.

Can we start early?

Background reading, team formation, account setup and access testing with designated practice examples are allowed in advance. Competition solutions, challenge-specific model training and judged deliverables begin Saturday at 10:30 a.m.

Where do we work overnight?

The overnight workspace is at the Courtyard by Marriott Sacramento Midtown, 4422 Y St, Sacramento, CA 95817. It is open from Saturday at 3 p.m. through Sunday submissions at 9 a.m., followed by pack-up and return to Aggie Square from 9–10 a.m. Participants arrange their own transportation. Scheduled on-site and remote mentoring will be coordinated through Discord.

What do we submit?

A project and a demo or pitch video of no more than three minutes through Devpost by Sunday at 9 a.m.

When must we return Sunday?

Expected return is 10 a.m. Every team member must check in individually and be present by 11:30 a.m. to remain eligible for overall awards, and teams must remain through the awards ceremony. Emergency or accessibility exceptions require approval from the event lead.

What about prizes and Nemotron?

Award categories and prizes will be announced once confirmed. A separate NVIDIA Nemotron award is proposed, subject to confirmation. Nemotron use is optional for the overall competition.

What should I bring?

A laptop, charger, curiosity and willingness to experiment.

Can students from other institutions participate?

No. The event is open to UC Davis students only.

The Harvest Is Happening.

The question is whether you can see it from space.

Registration is open on Devpost. Form your team of 5–7 and register before the event.

Register on Devpost

200 student places. Teams of 5–7, formed before the event.