Data Science solutions for humans

Our company started life as a data science consultancy.

We learned, over the years, that to truly transform our clients’ capabilities our solutions needed to be embedded. That is why we developed expertise in data platforms, software engineering and cloud operations.

Our core belief is that data science exists to support humans, not to supercede them. When we deliver a model, we consider:

• How will its data be collected, prepared and validated?

• How can outputs be explained?

• How should human experts be involved in the process and what control do they need?

• Where should model outputs be sent so that they result in effectively executed decisions?

The field has advanced in leaps and bounds since we started Unai, but these basic questions have always remained at the heart of what we do.

Typical Library & Language Choices

• Natural Language Processing (NLP) with spaCy, gensim, huggingface

• MLOps with AWS Sagemaker

• Reporting and Panels witt Power BI, Shiny (R) or Dash (Python)

• Documented research with RMarkdown and Quarto

• Deep Learning with PyTorch

• Feature Engineering & Statistical Learning with pandas + sklearn (Python) or the TidyVerse (R)

• Stacked gradient boosting machines for supervised learning

• pymc and OR-tools for operations research

Project Spotlight: Data Science Solutions