Have someone that knows what you need before you even need to ask.
The inspiration for Mimir came from the need to streamline the management of knowledge and data, helping teams collaborate more effectively on complex technical projects. We wanted to create a solution that leverages automation and intelligence to simplify the process of accessing relevant information quickly, enabling better decision-making and smoother project execution.
Mimir is a comprehensive platform that centralizes technical documentation, project updates, and resources, providing personalized recommendations based on user roles and project needs. It integrates with various tools and platforms, offering real-time insights and automated notifications. With advanced search capabilities, it helps users quickly locate critical information, while also recommending resources based on ongoing projects.
Mimír uses RAPIDS cuML to accelerate the clustering of incident reports, providing up to 10-50x performance improvements for large datasets. Key machine learning components include:
- TF-IDF Vectorization: Converting text descriptions into numerical vectors
- K-means Clustering: Grouping similar incidents together
The integration is achieved through a Go-Python bridge backends/database/clustering_cuml_bridge.go, which will later on be swapped for gRPC calls.
Using Go, Next.js, Python, cuML, React, Typescript, WeviateDB.
