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:
Agentic workflow for RecSys enhancement for MLE's.
A local-first conversational shopping agent that tracks changing intent and finds the right product earlier.
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.
A shopping copilot that knows when not to answer.
An autonomous ML-research agent: +0.0040 over baseline, 45 runs, $2.55, zero interventions
An adaptive shopping agent that turns evolving conversations into structured intent and asks the question that reduces uncertainty fastest.
Give it a benchmark and walk away. An agent that does the part of ML research nobody enjoys: 40 experiments at 2am. It found a feature we'd missed, and stopped itself when it stopped learning.
A conversational shopping agent over a frozen 50k-item Amazon catalog: hybrid BM25/dense/metadata retrieval, cross-encoder rerank, a calibrated dialog state machine, optional LLM rerank, 10-turn cap.
Ele is a robust AI-generated image detector built to spot synthetic media even after it has been degraded by real-world manipulations like JPEG compression, blur, resizing, and noise.
TrueSight detects AI-generated and manipulated images using provenance, visual reasoning, and robust image classification, even after compression, cropping, resizing, or reposting.
An autonomous ML research agent that proposes hypotheses, edits recommender code, trains and evaluates candidates, and learns from signed results with zero manual intervention.
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
RobustFusion detects AI-generated images reliably—even after real-world edits and transformations.
an AI shopping agent that asks useful follow-up questions and recommends the customer's hidden target product within at most 10 turns
Ten allowed, 2.2 used. Ten Turns treats this as recognition, not search: the shopper is quoting the target's own catalog row. HR@10 1.000, 54s, standard library only, bit-for-bit reproducible.
A universal autonomous RecSys research agent with adversarial peer-review & 4-pillar taxonomy that iteratively invents, optimizes, and converges PyTorch models.
RobustFake detects AI-generated images—even after resizing, compression, cropping, blur, and reposting.
An autonomous recommender research agent that safely proposes, tests, and audits model improvements inside an isolated sandbox.
Agent AI-yoh is an autonomous AutoML research agent that designs, runs, and audits recommender-system experiments, helping teams discover better models with reproducible, trustworthy evidence.
From oops to SOTA—an autonomous ML scientist that turns failed experiments, weird metrics, and fresh research into better models, one loop at a time.
Hazelnut is a conversational shopping agent, given a short customer message, the agent returns ranked catalog recommendations, surfacing the target product as early and as highly ranked as possible.
Shoppers don't describe products, they quote them. Copaon matches the quote back to the catalog: 50,000 products narrowed to one in two turns.
Your shopping assistant forgets what you said five minutes ago. Larpers Go Shopping fixes that, tracking evolving intent while never losing sight of what you're actually shopping for.
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.