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
A conversational shopping agent that stops ranking and starts identifying. Hit@10 1.0000, score 0.9531 against the baseline's 0.1067
Detecting AI-generated images with confidence, calibration, and robustness.
RunProof turns every AI Agent run into a persistent, privacy-safe receipt, helping teams explain failures, cancellations, latency and token usage without exposing sensitive content.
Compress it. Crop it. Blur it. Still caught.
A conversational shopping copilot that understands vague requests, remembers evolving preferences and delivers accurate recommendations quickly—without paid APIs.
Lightweight scam-detection middleware for e-commerce chat/agent pipelines , intercepts messages, returns an explainable risk score in milliseconds, no ML training required.
An autonomous machine learning agent that iteratively improves video recommendations beyond the official KuaiRand baseline.
An autonomous ML research agent that finds weaknesses in a recommender system, forms hypotheses, runs controlled experiments, and learns from both successful and failed results.
An offline shopping copilot that asks high-value questions to find a customer's hidden product in fewer turns.
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.
An autonomous ML research agent that builds and improves recommender models with best-first search, turning routine experimentation into a reproducible, clear workflow with minimal human intervention.
An autonomous ML researcher that searches over executable hypotheses, one measured code change at a time.
Find what fits, conversation by conversation.
Most detectors break under real-world corruption. Persist estimates corruption severity and calibrates its confidence accordingly — validated across generators, resolutions, and 15 transform types.
My submission for the 2026 TikTok Techjam "4. Shopping Copilot: AI Conversational Search and Recommendations". This was mostly generated by LLMs, aiming to understand how LLMs fare at such a problem
Most shopping assistants ask what you want. Ours works out what it doesn't know and asks that instead.
0.863 against a 0.107 baseline. 92.5% of a ceiling we measured. No model, no network, no dependencies.
An answerability-aware, evidence-grounded shopping agent that knows what to ask, how to rank, and when to commit.
When our agent broke, the log just said "error". Model? Tool? Sandbox? We built Oculith so every run leaves a privacy-safe trace that points at the exact layer that failed, not a black box.
A robust AI-generated image detector designed to generalise across unseen generators and remain reliable under real-world image transformations.
Welcome to modern team vibecoding: everyone's LLM is writing their slice of code right now. We're your new safe and effective editing layer, delivering one clean repo instead of a sea of conflicts.
Distraction is a custom implementation of the transformer layer optimised for local hardware. Running up to 46× faster than the baseline.
ASTRA is a shopping agent that routes buyers and browsers down separate retrieval paths, asks questions with highest information gain, gates recommendations on confidence and longterm memory relevance
We are building a distortion-resilient AI-generated image detector that uncovers synthetic artifacts even after heavy social-media compression, blur, noise, resizing and more.
337 – 360 of 614
