Develop python APIs using FastAPI framework, based on TDD process with pytest and behave, highly integrated with AWS services.Develop machine learning end-to-end pipelines using TFX. Build a custom TFX component to manipulate data and ingest time series into the pipeline. Implement a custom class-based solution to handle different types of data and create the neural network that best suits for each case. Create scripts to deploy a TFX pipeline on kubleflow and automatically select the best kubernetes-pod to run based on the task requirements.

Build an OLS (Ontology Learning System) that learns from different domains and suggests similar specific terms, not just based on simple synonyms. Train NLP models to recognize entities and predict categories over a whole document. Train model to learn word associations, create vector representations and obtain similarities between words from specific domains. Develop scripts to run every task containerized on docker, from data processing using multithreading to train models using GPUs. Create an API to consume this OLS as a service.