An example of how a tinder api may be structured and implemented. This solution uses WebAPI 2, Owin Startup, and Ninject for Dependency Injection (IOC container).

1. GET api/discoveries

• gets a list of discovered users to like/dislike.
• note: for a production environment, use either paged results or single user per request.

2. POST api/discoveries/like/{userId}

• likes the specified user.

3. POST api/discoveries/dislike/{userId}

• dislikes the specified user.

4. GET api/discoverysettings

• gets the current users discovery settings.

5. PUT api/discoverysettings

• updates the current users discovery settings.

6. GET api/matches

• provides a list of matched users.

7. GET api/users/{userId}

• provides the users profile information.

1. Replaced JSON.NET Serializer with Jil (https://github.com/kevin-montrose/Jil)
2. Deflate compression
3. Use of Singleton Pattern throughout code (using Ninject).
4. TPL usage throughout code (async/await).

1. The use of REDIS or similar for caching.
2. Use of an Enterprise service bus for cross-system notifications/push notifications.

1. This is not a full implementation. Things like the database interaction implementation were left out for simplicity sake.
2. This code assumes an authentication platform is also included.

• Ex: using an OAuth2 implementation that provides an access token with all requests on the authorization header.

3. Database implementation emulates a database interaction with a Task.Delay(50);.
4. All requests currently ignore any queryable data passed with the request for simplicity sake.
5. A simple console application test app was written and tested with the /api/discoveries method to test performance.

• Current tests show about 850/calls per second.
• System specs include Intel Core I7 3930K, 16GB DDR3 RAM, executing on a 250GB SSD.