About the challenge:

Modern development frameworks and scalable tools have fundamentally redefined our relationship with technology. The real value now lies in how we design, optimize, and deploy robust systems from the ground up to solve complex global problems, bridging the gap between a theoretical concept and software running in production.

Global Innovation Build Challenge (GIBC) V2 is built around that exact challenge. Organized as a premier, fully online, student-led international hackathon hosted by Kang Chiao International School (KCIS), Xiugang Campus, this iteration follows the benchmark established by GIBC V1, which engaged 552 global participants and secured prominent industry backing from sponsors like NordVPN and Saily. GIBC V2 offers an advanced engineering platform where student developers, researchers, and innovators design high-impact computational technology.

The three challenge tracks are structured around real-world industry problems that require technical autonomy, structural optimization, and domain-specific integration: foundational model architecture design, applied data engineering for critical industries, and unrestricted general invention.

Throughout the 10-week hacking period spanning July 11 to September 21, 2026, participants will have the flexibility to register, form teams of up to six members, and submit their solutions. All projects will be evaluated on an international stage by an elite panel of 10–15 industry professionals and expert judges, culminating in a virtual closing award ceremony on September 28, 2026.

Track 1: TECH (Foundational LLM Development)

Build and train a Large Language Model (LLM) capped strictly at 50 million parameters to push transformer architectures to their absolute limits. Your solution should include a custom-designed network architecture, data curation pipelines, and training methodologies developed entirely from scratch. We want to see how much language intelligence, processing efficiency, and task-specific accuracy your system can yield within strict small-scale computational boundaries.

Potential Applications: a domain-specific analytical assistant tailored for specialized legal or cryptographic discovery; a hyper-efficient language model highly optimized for local edge-device deployment; or a lightweight generative text engine trained to synthesize highly technical, niche datasets without relying on commercial API frameworks.

Track 2: Applied (Medical Technology & Financial)

Build a data-driven solution that leverages empirical, real-world datasets to resolve critical enterprise and societal challenges. Your solution must implement predictive machine learning models, diagnostic pipelines, or automated fintech architectures to streamline complex decision-making processes. We want to see robust data engineering that transforms raw, unstructured information into highly precise, actionable outcomes within the healthcare or financial sectors.

Potential Applications: biomedical device integrations that process live telemetry for diagnostic software; data pipelines for pharmaceutical research analytics; quantitative algorithmic trading models that execute strategies based on market indicators; or automated risk management and fraud-mitigation systems designed for corporate financial portfolios.

Track 3: Open (General Technical Invention)

Create a functional prototype that demonstrates technical novelty, cross-disciplinary engineering, and total creative autonomy. Your goal is to show how software engineering, hardware design, interactive media, or generative algorithms can be synthesized to construct a novel product from scratch. This track imposes no architectural limits or resource constraints, encouraging developers to build the next major innovation outside conventional structural boundaries.

Potential Applications: integrate hardware and software into IoT-enabled smart infrastructure; introduce advanced productivity frameworks that redefine daily digital workflows; deploy proprietary generative AI tools; or deliver full-stack web applications designed to address systemic environmental or community challenges.

Get Started:

1. Register on Devpost (July 11 - September 26)

2. Assemble or form group (1 to 6 members)

Assemble your group or decide to build as solo developer. Once your group is set, choose one of the three distinct tracks that fits your or your team's vision.

3. Join the community (Info access)

Hop into the official GIBC V2 Discord server. This is where you can network with other global participants and stay updated on announcements throughout the 10-week marathon.

Requirements

What to Build:

Build a fully functional prototype, model, or technical system that directly aligns with one of our three challenge tracks: Track 01: TECH (LLM 50M Parameters), Track 02: Applied (Med/Finance), or Track 03: Open (General Invention).

Depending on your track, your project must meet these specific technical baseline requirements:

For Track 01: You must build and train a custom Large Language Model capped strictly at 50 million parameters from scratch.
For Track 02: You must construct a system built entirely around empirical, real-world data science, predictive machine learning, or automated financial/medical pipelines.
For Track 03: You must build a hardware schematic, advanced software stack, or an autonomous generative algorithm that breaks conventional structural boundaries.'

The Core Requirement: Strong submissions across all tracks must show a working prototype, feature a complete end-to-end flow, handle real-world complexity, and be documented clearly enough that another developer could understand, run, and build on top of your work.

What to Submit:

Your Devpost submission must include all six of the following requirements to be eligible for judging on the global stage:

01 | Project Description: A clear, written explanation on your Devpost project page detailing exactly what you built, the specific enterprise or societal problem it solves, and how it works under the hood.
02 | Source Code: A link to a public GitHub or GitLab repository containing all project files. Your repository must include a robust README file covering comprehensive setup instructions, prerequisites, and usage guidelines so the judges can run your solution.
03 | Demo Video: A 2–5 minute video uploaded to YouTube, Vimeo, or Youku (with the link included in your submission). The video must demonstrate your project in action, show the technical architecture running, and explain your methodology.
04 | Built With: A complete breakdown listing all technologies, frameworks, libraries, APIs, and hardware tools used in your project (e.g., custom transformer modules, specific datasets, or full-stack software components).
05 | Team Information: A list of all team members using their real full names. This is strictly mandatory for the verification process and the issuance of official GIBC V2 participation certificates.
06 | Screenshots: At least 3 high-quality screenshots or images uploaded to your project page showing your interface, data pipelines, training loss curves, or technical setup in action.

For Finalists: Upon the finalist announcement, top-tier teams will be required to prepare their solutions for the virtual closing award ceremony on September 28, 2026. Further technical validation details will be shared at that stage.

Hackathon Sponsors

Prizes

1 non-cash prize

TBA

1 winner

Prizes TBA

Judges

TBA
TBA

Judging Criteria

LLM(Perplexity & Accuracy)
Language modeling quality on held-out validation sets. Lower perplexity and higher benchmark accuracy score higher.
LLM(Reasoning Performance)
Performance on logic, math, and commonsense reasoning tasks. Evaluated through standardized benchmarks.
LLM(Training Efficiency)
Compute efficiency, training speed, and memory footprint. Efficient architectures and training strategies rewarded.
LLM(Innovation)
Novel architecture decisions, creative data strategies, or unique training approaches. Bonus for originality.
MED / FINANCE(Innovation & Impact)
How novel is your invention? Does it address a genuine medical or scientific need with measurable impact?
MED / FINANCE(Technical Feasibility)
Does your prototype actually work? Quality of design, materials, code, and engineering methodology.
MED / FINANCE(Scientific Rigor)
Quality of research, experimental design, data collection, and validation of your approach.
MED / FINANCE(Presentation)
Clarity of your demo, explanation of your invention, and ability to communicate its value to judges.
GENERAL INVENTION(Creativity)
Originality and boldness of the concept. How unique is the idea? Does it push creative boundaries?
GENERAL INVENTION(Execution)
Technical quality, polish, and usability. Does the prototype work well? Is the experience refined?
GENERAL INVENTION(Impact)
Potential to inspire, educate, or solve a real problem. Does the project leave an impression?
GENERAL INVENTION(Presentation)
Clarity and quality of the final demo. Can the team effectively communicate their vision?

Questions? Email the hackathon manager

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

Online	Public
1 non-cash prize	7 participants