Learn to design real machine learning systems with the help of several open-ended machine learning problems.

Become a Machine Learning Engineer

Learn to design real machine learning systems with the help of several open-ended machine learning problems commonly asked during interviews at big tech companies. You will start by identifying the problem statement, understanding scale and latency requirements and defining metrics. Then you will come up with the architecture, select models, gather data and finally execute and evaluate the models offline and online.

Grokking the Machine Learning Interview

The feed-ranking system aims to maximize user engagement. So, let's start by looking at all the user actions on a Twitter feed.

# User actions
The following are some of the actions that the user will perform on their tweet, categorized as positive and negative actions. 

**Positive user actions:**

- Time spent viewing the tweet

- Liking a Tweet

- Retweeting

- Commenting on a Tweet  
	
**Negative user actions:**

- Hiding a Tweet  

- Reporting Tweets as inappropriate


Now, you need to look at these different forms of engagements on the feed to see if your feed ranking system did a good job. 

> 📝 All of the metrics we are going to discuss can be used to determine the user engagement of the feed generated by our feed ranking system.


# User engagement metrics

The following illustration shows how *feed engagement data* generates useful statistics for gauging user engagement.

The above illustration shows different positive and negative engagements on a Twitter feed. Let's see which of these engagements would be a good one to target as your *overall system metric* to optimize for.

## Selecting feed optimization metric
An important thing to understand in selecting a topline is that it's scientific as well as a business-driven decision.

The business might want to focus on one aspect of user engagement. For instance, Twitter can decide that the Twitter community needs to engage more actively in a dialogue. So, the topline metric would be to focus more on the **number of comments** on the Tweets. If the average number of comments per user increases over time, it means that the feed system is helping the business objective. 

Similarly, Twitter might want to shift its focus to overall engagement. Then their objective will be to increase average overall engagement, i.e., **comments, likes, and retweets**. Alternatively, the business may require to optimize for the time spent on the application. In this case, **time spent on Twitter** will be the feed system metric. 


Learn about the metrics to capture the success of a feed-ranking system. 

Metrics

Learn about the metrics to capture the success of a feed-ranking system.