Hope you are doing great.
Both of these databases are from NoSQL family. Both are good in their respective use cases. You can integrate both MongoDB and Cassandra with Hadoop.
With Hadoop Developer, it will be always better to select any No-SQL Databases rather than selecting any other course (In this case Cloud Computing).
As Hadoop is data processing framework and No-SQL databases are used to store, process, and integrate with Applications to process analyse Big Data stored in Hadoop HDFS.
Please review the following comparison and use case for both to choose one of the database:
1) MongoDB and Hadoop are often used together to deliver a powerful "Big Data" solution for complex analytics and data processing. With the MongoDB-Hadoop connector, you can easily take your data from MongoDB and process it using Hadoop. In MongoDB-Hadoop integration, we can use Hadoop input/output for MapReduce jobs, running MapReduce jobs against static backup files, and using Pig to build data workflows using MongoDB.
This link covers all the 3 use-cases in detail which are suggested for Hadoop-MongoDB integration
2) Cassandra does provide native Hadoop support and can also be integrated with Hadoop. It provides support for Hive (a SQL data warehouse built on Hadoop map/reduce) and Pig (a Hadoop-specific analysis language that many think is a better fit for map/reduce workloads than SQL).
3) MongoDB is really simple, flexible. It is very easy for development, applying schema changes, querying. It is nearly priceless for NoSQL newcomers.
4) Cassandra is much more advanced. It requires more thinking about schema design and query patterns. Horizontal scalability is one of the best in NoSQL market. Cassandra has a advantage in write operations as well.
5) Many MongoDB users look for Cassandra in case of real performance and scalability reasons .
Please go through the below link:
(They list the similarities and differences between Cassandra and mongoDB along with their characteristics)