We're looking for a data scientist/ analytics person with strong developing skills.
What you get:
A PostgreSQL relations database hosted on heroku with roughly 1,2M rows. It includes 40.000 movies.
Each movie has attached relations tables about:
- Genre (40 categorical data)
- Budget (numerical with currency ID)
- Talents like cast or director (100k different talents)
- Country of production (70 categorical)
- Language (70 categorical)
- Keywords (15k categorical)
- Release date (date)
What we need is a item-based recommendation service, that allows us to recommend similar items (e.g. using k-nearest-neighbor) based on input from a user.
Example Search / Input might be for example:
- Genre: Drama, Action
- Country of production: Spanish
- Budget: 5MUSD
- Language: Spanish
- Keywords: teenager, car
See attached image of the Ruby Application.
What we're looking for are
1) a paginated list of relevant similar movies sorted by relevance which we can used to display a relevanced sorted list of movies in the attached page.
2) Plus: for one move or or an array we need the possibility to receive an indicator for each movie how similar an item is.
3) we would need to have the ability to define weights manually for each input factor (if this cannot be trained in another way)
4) use segments (defined e.g. by budget or genre) where we can have optimized weight definitions (if this cannot be defined in a smarter way)
5) build a "success" Indicator for Talents (e.g. Actor, Director, Producer) that allows us to find similar "successful" Talents based on this indicator. The Indicator is calculated from several numeric data representing success.
We would need to have a secure Json Webservice that connects your recommendation service with our Ruby App.
We're looking - at the moment - for a Bootstrap Implementation which then can be replaced by a more sophisticated solution (e.g. SAP Hana) at a later stage now.
Therefore for the moment we're looking for a fixed price for the described solution (or if needed - less capable solution) which is
- (IMPORTANT) deliverable within the next 20 days.
- Deliverable means integrated, tested in staging enviroment and hosted in the Cloud (Amazon, Heroku)
- managed via github
- processes any request in a maximum of 10 seconds
Potential Implementations Approaches
[url removed, login to view] might be an option
Pretty straight forward on Node.JS: [url removed, login to view] + [url removed, login to view]
Google Apps App Engine
more ideas ?
Long Term seek
Even it's now a fixed budget project due to limited resources right now, we're looking for a longterm collaboration with a communicative person who's open to join a international team of technology geeks, marketers and business managers who all run one startup.
Thanks for reading and any proper (!) application.
Talk to you soon :-)
9 freelancers are bidding on average €1476 for this job
6+ years of experience in machine learning and master degree holder in computer science. Expertise in R,WEKA, JAVA,PYTHON, Hadoop, MapReduce. Worked on many projects in machine learning.
hi I am good with recommendations algorithm. I am interested in this project. I want an opportunity to start free lancing. I can give my contribution for project. Hope we will work long term. Skype : [url removed, login to view]