For some, March may bring good luck, bad omens, or spring gales, but at Cohere, it means that an exciting new year of language AI innovation is well underway. One of the highlights in March this year was our joint partnership with Qdrant at the Multilingual Semantic Search Hackathon hosted by Lablab. Brilliant participants from all over the world, and with diverse technical backgrounds, came together to ideate, build, and showcase innovative ways to implement Cohere and Qdrant to solve real-world problems. During the event, hackers also had the opportunity to learn from our experts through insightful workshops, keynotes, and mentoring sessions.

Hackathon participants were challenged to build a solution, using the Cohere API and Qdrant’s vector similarity engine and vector database, that corresponded to one of five categories: Internal Knowledge Base Search, Legal Document Search, Forum Search, Customer Review, or Recommendations. Projects received additional points if they used Cohere’s multilingual semantic search capabilities or the Cohere Generate endpoint.

We’re always incredibly amazed to see what people do with Cohere’s API, especially over a short timespan. We got to review a wide range of use cases, and the overall event winner also won the Customer Review category! Huge congratulations go to Team AI Disruptor and their project, AI Brand Intel. Cheers to you!!

AI Brand Intel: An Overview of the Platform and Its Capabilities

The AI Brand Intel platform enables businesses to monitor and analyze mentions of their brand across social media and news sites, all from one dashboard. With language translation and content intelligence capabilities, the platform can process content over 100 languages and generate responses in the user’s preferred language. Businesses can also upload internal documentation, policies, rules, and regulations in multiple languages to feed chatbots and semantic search engines, helping to streamline content management and enhance customer experiences.

AI Brand Intel was built using Cohere’s multilingual embedding model together with the Cohere Embed, Classify, Summarize, and Detect Language endpoints. The platform uses the Qdrant vector database to generate and save embeddings, as well as a Flask-based API backend and a Bubble-based, no-code front-end. Learn more about how they built their platform in their GitHub repo.

Prizes and Recognition: Celebrating Team AI Disruptor's Success

As the overall event winner, Team AI Disruptor won big: $2,000 in cash plus $5,000 Cohere credits and $5,000 in Qdrant Cloud credits. They also get to have a virtual coffee with Nils Reimers, our Director of Machine Learning at Cohere. The category win also earned them an additional $500 in cash, as well as $2,000 Cohere credits, $2,000 Qdrant Cloud credits, a lablab.ai certificate, and online promotion. They’ll also soon be featured on our app examples page with a step-by-step look into how they built AI Brand Intel.

A huge round of applause to everyone who participated in this exciting event, and we hope you had fun exploring, learning, and networking with other developers. Keep up the great work!

Stay tuned for more Cohere-sponsored hackathons coming soon to the lablab.ai event lineup.

Last but not least. We can’t wait to see what you can build! Sign up for a free Cohere account and start building.