Devpost
Participate in our public hackathons
Devpost for Teams
Access your company's private hackathons
Grow your developer ecosystem and promote your platform
Drive innovation, collaboration, and retention within your organization
By use case
Blog
Insights into hackathon planning and participation
Customer stories
Inspiration from peers and other industry leaders
Planning guides
Best practices for planning online and in-person hackathons
Webinars & events
Upcoming events and on-demand recordings
Help desk
Common questions and support documentation
Sort by:
A zero-trust control plane for enterprise AI agents that asks only for evidence that can change the action — and checks before retrying uncertain actions.
Shopping copilot that finds the exact product in ~2 turns — TechnicalScore 0.9715 (9× baseline), $0 and 0 LLM tokens on the scored path. Learned + LLM layers handle real shopper wording.
An autonomous exception closer that blocks unsupported waste custody claims, requests exact corrective evidence, resumes safely, and releases an evidence-linked packet.
An agent that investigates production incidents, decides what's safe to fix, and acts — but never trusts its own opinion of risk.
A gateway every agent tool call must pass through: a fail-closed policy engine and a hash-chained decision log that make cross-region data leaks structurally impossible, not just filtered.
A control plane for AI agents that pays its own way on evidence.
ShopPal is a lightweight shopping copilot that understands messy human intent, combines lexical and semantic search, and finds the right product fast, with zero inference-time API calls.
(No LLM Needed!) Shopping Copilot: Context-Aware Conversational Product Search- Edsel Tan
Optimizing food rescue to better the whole community.
ExceptionOS autonomously handles that resolution process. It keeps working until the exception is resolved and the business outcome is verified.
A deterministic eligibility calculator holds veto power over an ADK agent chasing a parent's specialty-med prior-auth, so a wrong AI date never becomes a real phone script.
SOW-TaskMaster is a multi-agent system that orchestrates the full Statement of Work (SOW) lifecycle — from customer project request to final signature — with humans only touched at genuine decision.
Nomi is a personal AI agent that operates your computer, remembers how you work, and gets things done, not just chat, actual work.
Deterministic multi-agent alignment and resonance engine preventing cascading execution failures through mathematical telemetry and runtime gating
Nowadays everyone can build & ship. But how fast is your distribution? DemoDominator turns any URL into studio-grade product walkthroughs w/ synchronized voiceovers & rich Video SEO schema in 25 sec
Collective road awareness that turns vehicle telemetry into predictive driver intelligence.
The doctor gives the plan once. Sanad carries it until reality matches.
Our assistant logged that it failed, never what was asked — privacy by design. This agent recovers what buyers need from public sources, refusing anything it can't cite or lawfully use.
An autonomous recommender-system agent that improves KuaiRand-Pure ranking while logging every hypothesis, code change, metric, and recovery.
A background release-recovery agent that detects real failures, assembles Gemini capability crews, reassembles when a plan breaks, acts within policy, and proves the workload actually recovered.
Finds what you meant, not just what you typed — 7.7x the baseline score, zero LLM cost.
DINOv2 AI image detector tested on 19 real-world corruptions, not just clean data. Caught and fixed a data shortcut causing a fake score. Added FFT forensics, raised worst-case AUROC from 0.82 to 0.88
AutoMend detects backend failures, diagnoses root causes with Gemini, executes safe recovery actions on Cloud Run, verifies the fix, and autonomously records every incident.
Agents shouldn't just execute tasks but should operate.Agent OS is the zero-trust control plane that runs AI fleets like production infrastructure and autonomously recovers when agents die mid-mission