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 deterministic, privacy-first shopping agent that clarifies intent, remembers preferences, and finds the right product in fewer turns.
Problem Statement 2: Autonomous Machine Learning Research Agent for Recommender Systems. Autonomous Research -> Code -> Evaluate loop to beat recommender baseline. Without humans in the loop.
An autonomous ML research agent for recommender systems: forms its own hypotheses, writes the code, reads the metrics, and decides what to try next.
RatLab Shopping Copilot is a fully local, zero-token AI search agent that uses multi-turn state tracking and hybrid sparse/dense retrieval to accurately find products without expensive LLM APIs.
Every generator leaves a tell. We find it and show you where
Most detectors break under real-world corruption. Persist estimates corruption severity and calibrates its confidence accordingly — validated across generators, resolutions, and 15 transform types.
ShopCopilot is a shopping agent using hybrid search, constraint-aware ranking and smart questions. It handles changing intent safely and scored a technical score of 0.74—7× the Track 4 baseline.
ParLLiaMent is an autonomous ML research agent that automates the full iterative experimentation loop for recommender system benchmarks.
Autonomous AI research agent that discovers, tests, and optimises recommendation models and features on real-world user behaviour data
We built an autonomous ML engineer that forms hypotheses, writes and tests models, learns from failures, and iteratively improves recommender system
An LLM agent that autonomously proposes, codes, tests, and iterates on recsys models for KuaiRand-Pure — zero manual intervention, +0.0022 delta over baseline, full run-by-run transparency.
An autonomous research agent that challenges a scientific claim, runs the experiments that could disprove it, repairs the runs that break, and returns an evidence-backed verdict with every check shown
FlowState: An autonomous ML agentic workflow that diagnoses data, iterates pipeline code, and optimizes recommender systems end-to-end by competing against itself, to make machine learning easier.
TactileOS turns any screen - text, photos, code, math — into a real-time 32×32 Braille pin matrix you can feel, with matching audio and haptics.
Multiple models. One verdict. Stronger image verification
Turn months of satellite imagery into field priorities in minutes, not days. 1415 Agri uses Gemini's reasoning to show farmers where to inspect first and make better field decisions.
Detect AI-generated images and test confidence under real-world transformations such as compression, blur, resizing, noise, colour changes and cropping, all through an interactive Gradio demo.
A constraint-aware conversational search agent for TechJam 2026 Problem Statement 4. The agent turns vague or changing shopping requests into ranked products from a frozen 50,000-item Amazon catalog.
AI-image detection that survives the real internet. Compression takes off-the-shelf detectors from 71% to under 1% — we hold 83% worst-case, and every verdict carries a measured confidence grade.
ChevKan replaces cloud RAG by compressing 1.5M database rows into continuous math curves. A native C++17 engine uses 0µs register calculus to ground a local 0.5B LLM completely offline for $0.
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.
Collective road awareness that turns vehicle telemetry into predictive driver intelligence.
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
A safety-gated five-agent system that autonomously proposes, runs, evaluates, and remembers recommender experiments—without exposing the hidden test set.