Make anyone able to do leading-edge machine learning on their own problems, in the cloud. They keep control of their own data, but they can benefit-from and share insights about similarly posed problems.
Most industry engineers & analysts are much more knowledgeable about their field than any data scientist, they just need guidance about the way a datascientist would approach their problem. Also, when they work with a data scientist for short periods of time, the analysis that is done is often left behind with few ways for other people to understand the exact analysis that was done.
By guiding the user using an index papers articles and code snippets that are matched to their aims for the analysis and that are mapped to the structure of their data, they can use the best practice of the expert data scientists
What it does
A guided analysis platform for data: Pick your
explore new data: show trends and & correlation,
recommends good strategies from index of NIPS conference files and kaggle
& executes analysis
How we built it
All on AWS.
Challenges we ran into
Getting stable beakernotebook on a virtual machine that serves a number of clients
Accomplishments that we're proud of
Custom search index Automatic indexing of user files Simple user flow
What we learned
how to use a search, indexing & NLP api how to structure & parse code snippets for good comparison