Devpost

🎉 The results are in! A huge congratulations to all 3,074 participants who took part in Build with DataHub: The Agent Hackathon! Over the past few weeks, developers pushed the boundaries of what’s possible when AI agents are backed by real data context. From building autonomous data engineering agents and generating production code to protecting ML pipelines in production. Out of an incredible field of submissions, here are the winners!

𝗪𝗵𝗮𝘁 𝗰𝗮𝗻 𝗮𝗻 𝗮𝗴𝗲𝗻𝘁 𝗱𝗼 𝘄𝗵𝗲𝗻 𝗶𝘁 𝗵𝗮𝘀 𝗿𝗲𝗮𝗹 𝗰𝗼𝗻𝘁𝗲𝘅𝘁 𝗮𝗯𝗼𝘂𝘁 𝘆𝗼𝘂𝗿 𝗱𝗮𝘁𝗮? That's what our 𝘉𝘶𝘪𝘭𝘥 𝘸𝘪𝘵𝘩 𝘋𝘢𝘵𝘢𝘏𝘶𝘣: 𝘛𝘩𝘦 𝘈𝘨𝘦𝘯𝘵 𝘏𝘢𝘤𝘬𝘢𝘵𝘩𝘰𝘯 answered. The results are in and the winners are here! 🏆 𝗚𝗿𝗮𝗻𝗱 𝗽𝗿𝗶𝘇𝗲: Project Blackbox by Alex V. ◼️ Blackbox takes one plain-English sentence like “revenue just jumped, is this real?” and works backwards. It reads DataHub lineage to localize the change, proves the root cause against a machine-enforced evidence gate, repairs the transformation code, runs the full invariant suite, opens a pull request for a human, and writes the resolution back to the catalog as an incident. ✨ Challenge winners ✨ 𝗔𝗴𝗲𝗻𝘁𝘀 𝘁𝗵𝗮𝘁 𝗱𝗼 𝗿𝗲𝗮𝗹 𝘄𝗼𝗿𝗸: Hindsight by Germán Massello 👀 Hindsight is an on-call agent for data platforms. It walks DataHub lineage to find what broke a dashboard, ranks the downstream consumers by who is affected, and writes its findings back as tags, incident banners, and a postmortem document. 𝗠𝗲𝘁𝗮𝗱𝗮𝘁𝗮-𝗮𝘄𝗮𝗿𝗲 𝗰𝗼𝗱𝗲 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻: Schema-Drift Auto-Repair Agent by Yadnesh Salvi ⚒️ This agent uses DataHub’s column-level lineage to compute the true blast radius of a rename, retype, or drop, then rewrites the affected dbt models and Airflow DAGs and opens a pull request with the diff, generated tests, and a migration doc. It also decides what not to change, separating code that must be updated from downstream models already insulated by an alias created upstream, and reporting every skipped model with the reason. 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗠𝗟: Culprit by Jonathan Andrei 🚕 Culprit is a stack trace for model decay. It walks DataHub’s ML lineage from a degraded model back to the column that caused it, then prices the damage. It runs on millions of real New York City taxi trip records. When a new vendor entered the feed and one upstream column’s maximum value changed from six to seven, every freshness, volume, null-rate, and schema check stayed green while the model produced $90,322 of attributable error in a single month. 𝗢𝗽𝗲𝗻 & 𝗪𝗶𝗹𝗱𝗰𝗮𝗿𝗱: Paracelsus by Lutfiya Miller, PhD, DABT and Chris Müller ⚠️ Point a data quality tool at a catalog and it returns hundreds of problems ranked by severity. Paracelsus argues severity is the wrong axis and borrows from toxicology instead: risk is hazard times exposure times receptor sensitivity times uncertainty. Its central idea is that a successor count is not exposure, but successor use is. Paracelsus reads real lineage, query history, usage statistics, tags, and domains from DataHub to work out which findings a team should act on rather than which look alarming, while also being transparent about its own limits. ✨ Honorable mentions ✨ 🔸 ContextSeal - Mehmet Aydoğan 🔸 RippleProof - Muhammad Hassaan Javed Congratulations to our winners! We had more than 600 submissions and saw some incredible projects. Dive deep into the details: https://bit.ly/4doZiYv

We've partnered with Devpost to bring the Amazon Developer Hackathon to our network. 🚀 For the first time, Fire TV, Alexa+, Ring, and Bee are all open to outside developers in one global hackathon. Build a demo for a single product line or connect several into one experience. Four primary tracks to choose from: Fire TV, Alexa+, Bee (wearable AI), and Ring. There is also an AWS Builder mini challenge for anyone who wants to layer in Amazon Bedrock, SageMaker, AgentCore, or the Strands SDK on top of their submission, a great fit for students who have been building with us this semester. Up to $190K in cash and AWS credits is on the table, with track winners also getting a 1:1 meeting with the Amazon Developer team. New to building for Amazon devices? Documentation, live office hours, and AI prototyping tools are available to help you go from idea to submission. Register here (also linked in our Linktree): https://lnkd.in/g_utmXTv 💻 #AWSsbgl #StudentsAtAWS #StudentsBuildAWS #StudentsBuildOn

Tomorrow at 8:00pm ET, submissions close. This is it. The Agents for Humans Hackathon wraps up September 14 and we're in the final stretch. If you've been building, now is the time to polish and submit. If you haven't started, you still have time. Here's what's on the line: 🏆 $40K in cash prizes across three tracks. 💳 $50 in AWS credits for every participant who submits. 🌍 Open to all experience levels. Enter solo or with a team. Three tracks to choose from: 🏠 Everyday Agents: Automate your everyday life 💼 Professional Agents: Build an agent that handles the repetitive work 🤝 Good Neighbor Agents: Help a community org that needs it Your agent doesn't need to do everything. It just needs to solve one problem and show it working. Submit your project before it's too late. Link in comments.

The Nebius AI Builder Program is officially live! Which model, where to run inference, which tools actually hold up 😅 The time it takes to figure it all out can take longer than the building. Nebius is coming in hot with a free solution with: → $400+ in credits and discounts from Nebius, LangChain, Toloka, Tavily, and more → Working code and cookbooks → Nebius Academy certifications → Office hours with engineers to debug, tune, and get unblocked → Meetups, hackathons, and a builder Discord community Partners include NVIDIA AI, LangChain, Hugging Face, Cognition, OpenHands, Tavily, Toloka, Composio, Prime Intellect, MiniMax, Qwen, and more. If you're already building for the Nebius x NVIDIA Global AI Hackathon, this is a natural next step. The office hours alone could save real time before the October 30 deadline Check it out today → bit.ly/nebiusnvidia26bp Register for the hackathon → https://lnkd.in/ejdEcmVv

8,460 builders. Five tracks. $70K+ in prizes. That's the Global AI Hackathon Series with Qwen Cloud has officially wrapped! Thank you to Alibaba Cloud for a genuinely ambitious challenge. With five tracks, production-grade requirements, and a community that showed up to build at that level, this was truly a hackathon to see. And to every builder who submitted a project: congratulations! Check out the full project gallery (including honorable mentions and blog award winners) on Devpost today: https://lnkd.in/eWvnKSJD

Our Build Sessions have been a hit! If you missed our recent one with Nebius, check it out on our Youtube today: https://lnkd.in/e69UNYBv There's still time to register for the Nebius x NVIDIA Global AI Hackathon too. Start your build on Devpost: https://lnkd.in/ejdEcmVv

Check out some of the winning projects for OpenAI Build Week Hackathon and see what they shipped with Codex! An amazing effort from about 42,000 builders all over the world and a great collaboration with OpenAI Developers. Congratulations to everyone who built and shipped something! See all the projects in the project gallery today: https://lnkd.in/euaPbxcY

One of our hackathon judges, Sandhya Subramani, is going live tomorrow at 1pm ET. Building an agent from a blank file to production in one session. Perfect timing if you're working on your hackathon submission or need a push to get started. 5 days left. 👇