buildstartupinpublic.substack.com is a free newsletter about less often told but frequently suffered pain-points by Startup Founders. Subscribe to receive our articles directly delivered to your inbox
Vevesta is a Collective Wiki/Knowledge repository with tracking and collaboration capabilities for Machine learning(ML) projects. Following are the USPs :
1. Feature Dictionary in Vevesta: Features (i.e. inputs to ML algorithm) discovery where data scientists can search features relevant to their project. This reduces the discovery phase of relevant features for machine learning projects.
2. Technique/Algorithm dictionary in Vevesta: Data Scientists, based on work of previous data scientists, can search algorithms relevant to their problem statements. This reduces the steep learning curve the data scientists face while finding, learning and implementing new relevant algorithms.
3. VevestaX: Open source library for automated check in of features, notes, code, algorithm used by data scientists using minimal lines of code. This is from inside the environment they are most comfortable with (Jupyter notebook/Python)
For people viewing using mobile, here is the youtube link to understand what the tool does: https://youtu.be/QSwRQv70iy4 or please mail vevestax@vevesta.com for a demo of the tool.
Hoping to hear feedback on if this product is useful for data scientists/ML engineers. The website is optimized for web viewing at the moment. Please press the explore button to understand how the tool works. The tool is tightly integrated with VevestaX(open source library) which enables you to push experiments from jupyter notebook/pythons script to the tool, Vevesta. Eager to hear your thoughts. Thanks, Priyanka
VevestaX is an open source library to track ML experiments and features in an excel file using 2 lines of code. The library automatically does EDA for you on the given dataset.