Abstract: The present disclosure generally relates to generating a care and treatment plan for a patient using machine-learning techniques. An exemplary computer-implemented method comprises: receiving a plurality of care records; determining, based on a first machine learning model, that the plurality of care records are to be merged into an episode, wherein the episode is a predefined data representation comprising a predefined collection of data fields; identifying, from the plurality of care records, a plurality of attribute values corresponding to a particular attribute; determining, based on a second machine learning model, an aggregated attribute value based on the plurality of attribute values; updating, based on the aggregated attribute value, a data field of the episode; adding the episode to the patient profile, wherein the patient profile comprises a plurality of episodes associated with the patient.