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
TraceGuard helps spot AI-generated images—and tells you when its own answer might be unreliable after an image is compressed, resized, or reposted.
AgentGate gives every AI Agent a separate, scoped, time-bound and revocable capability, with backend policy enforcement, one-time human approval and a redacted audit trail.
A multi-turn shopping copilot that understands evolving intent and finds the right product faster.
Much like how crystals are clear, KRYXstal aims to be the transparent middleware that offers verifiable visibility, least-privilege access control, & complete traceability over multi-agent execution.
Spots AI-generated images even after the internet has chewed them up. Point it at a folder and it scores every file, offline, with the false-positive rate calibrated so real photos aren't accused.
From vague shopping intent to relevant, constraint-aware product recommendations—fast, explainable, and privacy-friendly.
Adds capability-based access control, protected-resource relays, risk scoring, and audit trails to make AI agent actions enforceable, transparent, and accountable.
An autonomous LLM agent that plays the role of an ML engineer on a recommender system: it hypothesises, codes, trains, evaluates, and iterates, but with no human touching the model in the loop.
Proof-carrying multi-agent decisions for worlds that actually existed.
A stateful shopping agent that asks smarter questions, adapts to changing intent, and ranks the right products quickly and reliably.
Optimize PyTorch transformer specifically for NVIDIA GeForce RTX 5060Ti 16GB card, targeting lower cuda latency and higher model FLOPs utilization (MFU).
Narrow is a stateful shopping copilot that turns evolving conversations into dynamic multi-route retrieval, safe personalization, and explainable Top-10 recommendations.
A shopping copilot that uncovers hidden needs from conversation, adapts when preferences change, and explains every recommendation.
Raiding the fakes real-world edits couldn't hide.
Agents can cook, just not outside their permissions. AgentFence enforces runtime boundaries and gives you receipts for every delegation, denial, and action.
An autonomous recommender-system agent that improves KuaiRand-Pure ranking while logging every hypothesis, code change, metric, and recovery.
Compression and reposting change an image's pixels but not its origin. We fine-tune a vision encoder so its features ignore that damage, and detection accuracy holds where it usually drops.
Autonomous Machine Learning Research Agent for Recommender Systems
R&D that doesn't sleep: an autonomous agent that improves your recommender overnight while you zzz
An AI-generated image detector that combines an RGB AIGC expert with low-level forensic signals such as FFT energy and pixel statistics, then dynamically fuses both under real-world transformations
RIFT detects AI images after compression, blur, resizing and reposting. It combines frozen CLIP with forensic signals and uses a fixed low-FPR threshold for realistic social-media evaluation.
Your agent says the run completed. The trace says which step actually failed. Glass Box middleware: a span waterfall per Agent Run, plus a policy gate that denies secret exfiltration.
Agents get capabilities, not credentials
Seeing through the fog of AI 😎
97 – 120 of 614
