Machine Learning Ops or MLOps defines a systematic approach of building, deploying and maintaining Machine Learning solutions. It encompasses the practical skills and techniques required to deploy machine learning models into production and building a retraining pipeline with the inference data to retrain the model in order to resolve the data-/concept-drift issues.
ML Engineering for production combines a basic understanding of Machine Learning and experise in Applied Software Engineering and DevOps.
Read MoreThis section talks about building a datapipeline by collecting, cleaning, pre-processing, validating and assesment of the datasets, and also of feature engineering, and labelling datasets.
This section talks about tooling and techniques for managing various model artifacts, model's performance, audit system and model's interpretability.
Read MoreBuilding a scalable and resilient hardware architecture to cater to real-time and batch inference requests. This section describes different patterns of model serving patterns and infrastructures.
Read MoreMachine Learning projects requires scaling and operational efficiency when deployed in production. This section focuses on building, training deploying and monitoring Machine Learning and AI products on cloud. This section also describes how to build end-to-end automated and human in loop ML pipelines in AWS Cloud and training and deploying BERT for state-of-the-art NLP and NLU on AWS Cloud.
Read More