Build an infrastructure to serve a pre-trained machine learning model.
Customers will send raw data to the service.
This raw data will be processed and scored using the model.
The score will be sent back to the customers.
The raw data must be stored for future use (For example model improvement, not going to be implemented during this project).
The goal is to build a scalable infrastructure that adjust the used resources depending on the load.
Technologies to be used:
Containers:
Docker
Orchestrator:
Kubernetes
Mesos
Message broker:
Kafka
Redis
RabbitMQ
Storage:
Cassandra
MangoDB
Redis
etc
Processing engine:
Spark
Flink
Storm
The goal is to understand the advantages and disadvantages of these technologies. Each member of the team should research the different offerings in one category [Orchestrator, Message broker, Storage, Processing engine] and determine which technology is most convenient to use in our application.
Project Idea:
Build an infrastructure to serve a pre-trained machine learning model.
The goal is to build a scalable infrastructure that adjust the used resources depending on the load.
Technologies to be used: Containers:
Orchestrator:
Message broker:
Storage:
Processing engine:
The goal is to understand the advantages and disadvantages of these technologies. Each member of the team should research the different offerings in one category [Orchestrator, Message broker, Storage, Processing engine] and determine which technology is most convenient to use in our application.