Enolink is a startup that provides a data science platform tailored to healthcare use cases, which includes building their own AI/ML models. The company supports federated learning and collaboration infrastructure for training models on large, heterogeneous datasets, and facilitates multi-endpoint analytics from simple queries to AI/ML modeling.
Startups building their own models
Deepgram is a startup that is involved in building their own models. The company is part of the Y Combinator W16 batch, indicating its startup status. Additionally, Deepgram provides advanced speech AI transcription and understanding via an API, and allows for the training of custom models for unique use cases, which shows that they are involved in building their own models.
Activeloop is a startup involved in building their own models. The company is a small team (10 people) and is backed by Y Combinator, indicating its startup status. Additionally, the company provides a database for AI and offers tools for creating, storing, versioning, and collaborating on multi-modal AI datasets, which implies they are involved in building their own models.
Shuimu BioSciences is a startup founded in 2017 and is involved in building their own models, specifically through their development of the modelSMART deep learning algorithm for atomic structure modeling.
AssemblyAI is a startup that has raised significant funding and is involved in building their own AI models for various tasks such as audio transcription, content moderation, and summarization. They have developed models like Conformer-2 and are working on next-gen models leveraging advanced technologies.
Cellworks is a startup involved in personalized therapy biosimulation for cancer treatment. They utilize a biosimulation platform to predict patient responses to therapies, which indicates they are building their own models.
Aurora is a startup involved in developing autonomous driving technology, which includes building their own models as evidenced by their Data Engine that accelerates machine learning model workflows.