Nerea Luis – Artificial Intelligence seen from the software development lifecycle perspective

Testing Ai

Artificial Intelligence development requires robust infrastructure, collaboration, and a systematic approach with a focus on transparent, interpretable, and maintainable models.

Key takeaways

Artificial Intelligence development is similar to software development; both require a robust infrastructure and collaboration between teams.
AI models are not just a piece of software, but a living thing that requires attention, testing, and validation.
Nerea Luis emphasizes the importance of having a proper data governance policy in place, especially when it comes to AI models.
She also highlights the need for automating the way AI models are deployed and monitored, as well as automating data cleaning and logging.
The importance of understanding your data culture and moving towards a data-driven approach was repeatedly stressed.
The concept of MLOps (Machine Learning Operations) was introduced, emphasizing the need for a more systematic approach to AI development.
Nerea Luis says that AI models are not as easily interpretable as they seem, and that the onus is on the developer to ensure the model makes sense and is accurate.
The role of software developers in AI development was discussed, highlighting the need for cultural awareness and collaboration.
The speaker emphasizes the value of testing and validation in AI development, as well as the need to consider factors like data quality and bias.
The talk also touched on the concept of explainability in AI, highlighting the need for transparency and accountability.
Nerea Luis concludes by stating that AI models are not magic, and that they require careful development, testing, and maintenance.

Nerea Luis – Artificial Intelligence seen from the software development lifecycle perspective

More talks