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

# Architecture
Let's have a look at the architectural components that are integral in creating our Twitter feed system.  

Let's briefly look at each component here. Further explanation will be provided in the following lessons.

## Tweet selection
This component performs the first step in generating a user's feed, i.e., it fetches a pool of Tweets from the user's network (the followees), since their last login. This pool of Tweets is then forwarded to the ranker component. 

## Training data generation

Each user engagement action on the Twitter feed will generate positive and negative training examples for the user engagement prediction models.

## Ranker
The ranker component will receive the pool of selected Tweets and predict their probability of engagement. The Tweets will then be ranked according to their predicted engagement probabilities for display on user A's feed. If we zoom into this component, we can:

1. Train a single model to predict the overall engagement on the tweet.
2. Train *separate models*. Each model can *focus on predicting the occurrence probability of a certain user action for the tweet*. There will be a separate predictor for *like*, *comment*, *time spent*, *share*, *hide*, and *report*.  The results of these models can be merged, each having a different weight/importance, to generate a rank score. The Tweets will then be ranked according to this score.

   The following illustration gives an idea of how the ranker component combines the outputs of these models to come up with a rank score for a Tweet.

Have a look at the architectural components of the feed ranking system.

Before We Begin

Introduction

Practical ML Techniques/Concepts

Search Ranking

Feed Based System

Recommendation System

Self-Driving Car: Image Segmentation

Entity Linking System

Ad Prediction System

Conclusion

Architectural Components

Architecture

Tweet selection