# AssemblyAI Winter Hackathon Winners ## Метаданные - **Канал:** AssemblyAI - **YouTube:** https://www.youtube.com/watch?v=rve3ioH-004 - **Дата:** 15.12.2022 - **Длительность:** 9:05 - **Просмотры:** 640 ## Описание We had an amazing weekend with everyone working on their own AI-first product. So many amazing projects were made in such a short time! Here are the winners of the Hackathon! ▬▬▬▬▬▬▬▬▬▬▬▬ CONNECT ▬▬▬▬▬▬▬▬▬▬▬▬ 🖥️ Website: https://www.assemblyai.com 🐦 Twitter: https://twitter.com/AssemblyAI 🦾 Discord: https://discord.gg/Cd8MyVJAXd ▶️ Subscribe: https://www.youtube.com/c/AssemblyAI?sub_confirmation=1 🔥 We're hiring! Check our open roles: https://www.assemblyai.com/careers ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ #MachineLearning #DeepLearning ## Содержание ### [0:00](https://www.youtube.com/watch?v=rve3ioH-004) welcome everyone to the recap video of the AI winter hackathon that assembly I hosted last weekend it was an incredible event we had 440 participants from 84 different countries and more than 150 submissions in the end and the goal was to build an AI powered product but you were free to choose the domain and also the Technologies you wanted to use and it was really incredible to see what you could all came up with all the different ideas and what you could all accomplish in just two days like seriously we were speechless when the projects came in on Sunday and it was really hard to choose a winner so I want to thank you all for spending this weekend with us for putting in the time and the effort and building I hope you learned a lot and also enjoyed this we sure did and maybe we will see you in the next one and now in this video we want to go through all the winners and all the honorable mentions I'm sorry that not all projects could make it to the finalist like I said it was really hard to choose a winner but in the end we had to make a decision so here are our top picks and now let's go over them in the first place we have super paint by Assad Memon ### [1:11](https://www.youtube.com/watch?v=rve3ioH-004&t=71s) Superpaint and Asuka Kaiser super paint is a drawing app that can enhance your sketches with AI so given that initial painting you can generate variations of your painting by giving a description and a specific style you want your variations to be in you can generate parts of your drawing by highlighting a certain parts and providing some context and you can also expand your drawing and generate a framing for it finally it is also possible to in-paint objects to existing images this app was built with react and another technology called fabric. js and the image technology is coming from stable diffusions image to image model and in painting and out painting models that are on replicate we really like this project for how interactive it is and on top of that all the features are already built to an MD EP level and the interface is very user friendly and intuitive and on top of that the app was very well presented so congratulations to the team for buildings project and taking home the 35 000 first place prize the second place goes to Toy Story Creator by sahit and ### [2:18](https://www.youtube.com/watch?v=rve3ioH-004&t=138s) Toy Story Creator tensei this brings children's toys to life for bedtime stories so here the user can upload images of toys and then select the age for which those stories should be created and then requests are sent to the cohere API to create the stories plus requests are sent to replicate to train a model and then the train model can be used to create new images that go along with the story and then everything is combined into one bedtime story and the interface was really easy to use and both the landing page and the final story Page look really beautiful we also left the complexity of combining multiple Technologies and apis plus the huge potential to build on top of this in the future so you listed some possible improvements on your submission for example you could combine this with assembly I speech to text API to allow children to dictate their own story or to implement Uber deck to read the stories so there is really a huge potential to improve this in the future so great job saheed and tensei and congrats on winning second place and the ten thousand dollars plus one thousand in API credits great job the third place went to operator AI that was made by ### [3:36](https://www.youtube.com/watch?v=rve3ioH-004&t=216s) OperatorAl damir Tamir Naveen ravindran lyric operator AI is an AI system that connects 911 callers with their required Emergency Services the team came up with this idea during the recent hurricane in Florida when they saw that there were not enough emergency services available to everyone who needed help with operator AI they are aiming for the calls to be dealt with quickly and effectively and helping to locate people faster so that you can start dispatching help to them faster how the app works is once an emergency number is called the caller is guided to give their names their location and the type of emergency they're facing the location and the emergency type is extracted from the caller's answer and a priority level is automatically assigned a dispatcher can later update these values if they see fit one of the nice things is that this app was built using a lot of technology so first of all they use civilio to automatically answer calls and guide the callers and then assembly AI to real-time transcript what the caller is saying and then name entity recognition and text classification models from hugging phase to understand what the user is saying and to be able to assign it a priority level the front end of the app is made by react and Tucker UI Y and the backhand we have digital ocean and Firebase we loved how thoroughly this app was prepared and how multiple Technologies from different sources were combined to create an app that would actually be useful in a real life setting it is also very thoughtfully made because it really covers the important pain points of the problem that they're trying to solve and it was very nicely presented at the end so congratulations to the team for winning the third place prize of five thousand dollars plus 500 and assemble the AI API credits now we have five honorable mentions that will receive swag and also will be mentioned on our blog post so we have super cut by Zack and Xavier ### [5:39](https://www.youtube.com/watch?v=rve3ioH-004&t=339s) Supercut Supercut is an autonomous podcast editing assistant that automatically creates short social video clips and tweets from your podcasts video recording then we have photo studio by a seam that lets you create stunning photo ### [5:52](https://www.youtube.com/watch?v=rve3ioH-004&t=352s) PhotoStudio shoots for your products in seconds by using AI to segment the product and then applying stable diffusion in painting to create new images with the product then we have scaling Potato by Danny Lisa and ### [6:05](https://www.youtube.com/watch?v=rve3ioH-004&t=365s) Scaling Potato animesh scaling potato is an AI power technical interview preparation platform where users can practice a mock interview with an AI guided interviewer then we have may book by Anushka and Mac a fun and personalized learning platform for kids it uses AI to create books with comic illustrations to take the kids on an adventure while at the same time teaching about the topic and the last honorable mention will go to Braden and David's project which is an AI power journaling product it uses AI to transcribe your spoken thoughts into a journal entry and generate a complementary image that captures its Essence on top of the first second and third place prizes and honorary mentions and the hackathon we gave away two special prizes one of them is the assembly AI special price and the other one is coherence special price so as you can understand from the names the assembly AI special prize was given to a project that used assembly AI in a very creative way and the cohere special price was given to a project that used cohere in a creative way so the assembly AI special prize went to pupil. ai is an AI assistant to ### [7:18](https://www.youtube.com/watch?v=rve3ioH-004&t=438s) Pupil.ai replicate asking questions to a professor during a lecture in an online learning setting with pupil. ai's Chrome extensions Learners can ask questions to a chatbot about the contents of a video and in a time where we do a lot of our learnings online you can really see how useful this app can be even day to day we really love the creative use case this team came up with and how they used assembly AI to make online learning more interactive and the special cohere prize went to Doc Space by Nathan dalberg ### [7:50](https://www.youtube.com/watch?v=rve3ioH-004&t=470s) Docspace docspace is a legal document analysis tool that lets users extract specific information from complicated legal documents given the PDF of a complaint it extracts each claim in this case making it easier to identify whether a motion was fully or partially granted or denied eliminating the need to dive into the original documentation to retrieve you this information we loved how using models from cohere in a creative way Nathan was able to create a novel and very useful document analysis tool so great job everyone and congratulations again we really had a lot of fun going through all those projects and I hope that you don't stop here but improve those in the future even the ones that didn't make it to this list there are still a lot of other ones that were really great and that have a huge potential to be used as an actual product so work more on them and launch them till the public this would be really cool to see and yeah I hope you enjoyed this weekend and maybe join us for the next one and now go to the dev post page and there you can see the project Gallery so take a look at all of the other projects as well and again congratulations to the winners and then I hope to see you maybe in the next one bye --- *Источник: https://ekstraktznaniy.ru/video/12868*