It's a mixed-format engineering challenge for builders who want their agentic systems to work in real-world conditions. You team up to design and implement an agentic platform built for real use. The goal is a system that operates reliably in practice.
A structured challenge brief, evaluation criteria, and production-oriented requirements.
Live sessions and chat communication with R&D experts throughout the hackathon.
Get a prompt that demands speed and engineering discipline.
With your team, build a usable experience, a resilient platform, and an end-to-end agent workflow that integrates with existing tools and runs reliably.
Context engineering
Observability
Smart orchestration across tools, flows, and automation
Clear documentation of key decisions and outcomes
AI security
Unlock R&D career paths with direct visibility to hiring decision-makers
Exposure to advanced agent orchestration practices
Real-life AI engineering use cases
Opportunity to connect with R&D experts
Unlock R&D career paths with direct visibility to hiring decision-makers
Real-life AI engineering use cases
Exposure to advanced agent orchestration practices
Opportunity to connect with R&D experts
Agentic AI Engineer
Associate Director of R&D Vision Practice
Senior Agentic AI Engineer
Agentic AI Engineering Lead
Agentic AI Lead
Associate Director of R&D Products
Intelligence Engineering Lead
*Mentors aren’t assigned to specific teams. This keeps access equal and fair feedback for everyone.
Please note that all projects developed during the hackathon must be open sourced under the MIT license.
This ensures transperancy, shared learning, and equal opportunity for others to build on theoutcomes.
Laptop with Git + IDE (VS Code acceptable)
Node.js or Python runtime
Access to GenAI coding tools (Copilot, Cursor, ChatGPT, Claude Code, etc.)
API key for an LLM (multimodal preferred)
Secure handling of secrets (environment variables only)
REST / webhook integrations (real or mocked adapters)
Small UI or API intake
Multi-step agent flow with error handling
Structured logs and correlation IDs
Basic prompt-injection defenses
Readme
Demo video
Brief architecture and outcomes notes
Each participant registers individually. If you plan to participate as a team (1–4 members), simply enter the same team name in your registration form — we will use this to group you into teams. An expert committee will evaluate the acceptance of each registered participant, and we will notify them of acceptance or rejection of participation.
Get the challenge, your toolset, and step into the Discord workspace with mentors and other teams.
Work with your teammates, join expert sessions, ask questions, explore ideas, and shape your solution.
Share your video demo, deck, and code. The top teams move forward — show us what you’ve built.
The top 10 teams earn a spot in the final — and if you're among them, you’ll have to prepare for the pitch.
Enjoy the final event, present your work on stage, listen to others, network, and celebrate the winners.
Ship agents without guesswork. Learn how to monitor, trace, and evaluate AI systems in real environments so you can spot failures early, measure performance, and improve reliably. After this session, discover how to make your AI observable, accountable, and production-ready.
Turn your bold idea into something agents can truly deliver. After this session, learn how to move from a rough vibe to a clear plan and spec, writing constraints and acceptance criteria that drive real, testable outcomes. Walk away with a practical framework you can use immediately to build with confidence.
Test your engineering thinking in real conditions. We’re ready to support your next step.
Test your engineering thinking in real conditions. We’re ready to support your next step.
Here, research becomes deployed AI systems used in healthcare, enterprise, and global technology ecosystems.
See what drives R&D forward.
Here, research becomes deployed AI systems used in healthcare, enterprise, and global technology ecosystems.
See what drives R&D forward.
