At Londogard we find order in chaos by creating structures and understanding out of unstructured data. Our belief is that by creating understanding of data new possibilities open up and a lot of automation is possible where today manual tedious work is applied.
Our goal is to employ Machine Learning models that makes sense and provide value rather than being made to tick of a check-box. The aim is to provide Efficient, Performant, Measurable & Understandable models.
- Efficient should truly run on a single small machine (the edge)
- Performant achieving very close to State-of-the-Art performance, striking a good balance of performance & efficiency.
- Measurable in a way that makes sense for your personal use-case. Not some random metrics.
- Understandable to the level where you understand why a prediction was made as it was.
Through our blog & demos we try to show-case powerful models that all run on the same single Raspberry Pi 4 (4GB)