Three ways to challenge the system.

Problem

Authorization policies must account for roles, attributes and context while making security decisions understandable and testable.

Challenge

Build or evaluate an authorization capability around AeglysAI. Compare policy behavior under changing context, identify failure cases and explain the security tradeoffs with reproducible evidence.

Timeline

Planned launch: November 2026
Planned submission: January 2027

Exact deadlines, time zones and review dates will be announced in official terms.

Eligibility

Free participation for platform engineers, cloud engineers, security engineers, SREs, software engineers, AI/ML engineers, researchers and graduate students. Final eligibility, team limits and any geographic or age requirements will be specified in official event terms.

Deliverables

• Source code or evaluation harness and setup instructions
• Policy examples, threat assumptions and reproducible test cases
• Results with limitations and an explanation of authorization decisions

Proposed scoring framework

• Technical Correctness — 25%
• Security & Threat Reasoning — 20%
• Architecture & Engineering — 20%
• Reproducibility — 15%
• Innovation — 10%
• Documentation & Explainability — 10%

Independent judges score eligible submissions using the same evidence-based rubric. Conflict disclosure, recusal and a published tie procedure protect independent review. The framework remains subject to finalized official terms.

Prize

Planned USD $300 prize per hackathon. Prizes and rules are subject to official event terms until finalized. No prize has been awarded. Award conditions and payment details remain to be finalized.

Submission process

Required submission package (subject to finalized official terms):

• Public GitHub repository unless an exception is approved before submission
• README
• Architecture description
• Setup instructions
• Demo
• Test evidence
• Security considerations
• Limitations
• License information

Optional: Video demo, Benchmark results, Research notes. A demo is required; a video is optional. Submit the repository URL and a specific commit/tag through the official form when available. Receipt is separate from eligibility and judging.

Submissions are not open yet.

Rules

Proposed rules: use authorized isolated environments; avoid real credentials or personal data; disclose reused code, datasets and AI assistance; credit collaborators; and report results honestly.

Final eligibility, team size, permitted prior work, submission deadlines and award procedures are subject to official event terms.

Open-source expectations

A public GitHub repository is required unless an exception is approved before submission. Include source code, setup instructions, evaluation artifacts and explicit license information. Respect upstream licenses and document third-party dependencies and data provenance.

Use synthetic or appropriately licensed data; never publish secrets or sensitive datasets. Final licensing and submission requirements will be defined in official terms.

Problem

Behavioral signals can be noisy, and risk assessments need evidence that distinguishes suspicious activity from legitimate variation.

Challenge

Prototype or evaluate behavioral risk methods using synthetic or appropriately licensed data. Explain signal selection, assess false positives and document uncertainty. Model D is planned research, not an existing implemented capability.

Timeline

Planned launch: December 2026
Planned submission: February 2027

Exact deadlines, time zones and review dates will be announced in official terms.

Eligibility

Free participation for platform engineers, cloud engineers, security engineers, SREs, software engineers, AI/ML engineers, researchers and graduate students. Final eligibility, team limits and any geographic or age requirements will be specified in official event terms.

Deliverables

• Prototype or evaluation code with reproducible setup
• Data provenance, signal definitions and privacy considerations
• Evaluation method, false-positive analysis and explainable risk evidence

Proposed scoring framework

• Technical Correctness — 25%
• Security & Threat Reasoning — 20%
• Architecture & Engineering — 20%
• Reproducibility — 15%
• Innovation — 10%
• Documentation & Explainability — 10%

Independent judges score eligible submissions using the same evidence-based rubric. Conflict disclosure, recusal and a published tie procedure protect independent review. The framework remains subject to finalized official terms.

Prize

Planned USD $300 prize per hackathon. Prizes and rules are subject to official event terms until finalized. No prize has been awarded. Award conditions and payment details remain to be finalized.

Submission process

Required submission package (subject to finalized official terms):

• Public GitHub repository unless an exception is approved before submission
• README
• Architecture description
• Setup instructions
• Demo
• Test evidence
• Security considerations
• Limitations
• License information

Optional: Video demo, Benchmark results, Research notes. A demo is required; a video is optional. Submit the repository URL and a specific commit/tag through the official form when available. Receipt is separate from eligibility and judging.

Submissions are not open yet.

Rules

Proposed rules: use authorized isolated environments; avoid real credentials or personal data; disclose reused code, datasets and AI assistance; credit collaborators; and report results honestly.

Final eligibility, team size, permitted prior work, submission deadlines and award procedures are subject to official event terms.

Open-source expectations

A public GitHub repository is required unless an exception is approved before submission. Include source code, setup instructions, evaluation artifacts and explicit license information. Respect upstream licenses and document third-party dependencies and data provenance.

Use synthetic or appropriately licensed data; never publish secrets or sensitive datasets. Final licensing and submission requirements will be defined in official terms.

Problem

Distributed-system incidents require observable evidence and response mechanisms constrained by explicit security policies.

Challenge

Build or evaluate an incident-detection or resilience prototype in an isolated environment. Define response boundaries, test failure scenarios and explain recovery behavior. Autonomous response is a challenge theme, not a claim of deployed AeglysAI functionality.

Timeline

Planned launch: January 2027
Planned submission: March 2027

Exact deadlines, time zones and review dates will be announced in official terms.

Eligibility

Free participation for platform engineers, cloud engineers, security engineers, SREs, software engineers, AI/ML engineers, researchers and graduate students. Final eligibility, team limits and any geographic or age requirements will be specified in official event terms.

Deliverables

• Prototype or evaluation harness with architecture documentation
• Isolated incident scenarios, telemetry and reproducible runs
• Response policies, safeguards, recovery analysis and limitations

Proposed scoring framework

• Technical Correctness — 25%
• Security & Threat Reasoning — 20%
• Architecture & Engineering — 20%
• Reproducibility — 15%
• Innovation — 10%
• Documentation & Explainability — 10%

Independent judges score eligible submissions using the same evidence-based rubric. Conflict disclosure, recusal and a published tie procedure protect independent review. The framework remains subject to finalized official terms.

Prize

Planned USD $300 prize per hackathon. Prizes and rules are subject to official event terms until finalized. No prize has been awarded. Award conditions and payment details remain to be finalized.

Submission process

Required submission package (subject to finalized official terms):

• Public GitHub repository unless an exception is approved before submission
• README
• Architecture description
• Setup instructions
• Demo
• Test evidence
• Security considerations
• Limitations
• License information

Optional: Video demo, Benchmark results, Research notes. A demo is required; a video is optional. Submit the repository URL and a specific commit/tag through the official form when available. Receipt is separate from eligibility and judging.

Submissions are not open yet.

Rules

Proposed rules: use authorized isolated environments; avoid real credentials or personal data; disclose reused code, datasets and AI assistance; credit collaborators; and report results honestly.

Final eligibility, team size, permitted prior work, submission deadlines and award procedures are subject to official event terms.

Open-source expectations

A public GitHub repository is required unless an exception is approved before submission. Include source code, setup instructions, evaluation artifacts and explicit license information. Respect upstream licenses and document third-party dependencies and data provenance.

Use synthetic or appropriately licensed data; never publish secrets or sensitive datasets. Final licensing and submission requirements will be defined in official terms.