Introduction
In the bustling world of artificial intelligence, a persistent question remains: does machine learning engineering still have its place in a landscape dominated by generative AI? Vicki Boykis, during her talk at the Applied Machine Learning Conference in 2026, raised this issue by sharing her experience and thoughts on the evolution of our industry.
The Rise of Generative AI
Since the advent of language models like GPT-3 and its successors, generative AI has redefined the standards of technological innovation. According to a McKinsey study, the generative AI market is expected to reach $110.5 billion by 2030. This exponential growth raises the question of the relevance of traditional machine learning methods.
Machine Learning vs Generative AI
Machine learning engineering is based on solid fundamentals: data collection, cleaning, modeling, and evaluation. However, with LLMs (Large Language Models), these steps are often automated, simplifying the development process. For Boykis, the crucial question is: "Is it still worth doing machine learning well when the focus is on shipping quickly?"
Maintaining the Context Window
At a conference at PyData Amsterdam in 2024, Boykis emphasized the importance of "building and keeping your context window." In a context where tools are constantly evolving, understanding the history and underlying concepts of machine learning remains essential to effectively using new technologies.
Use Case: Real-Time Personalization
Boykis shared her experience at Malachyte, where she builds real-time personalization systems. In these cases, integrating generative AI with traditional techniques allows for optimized user experiences. For example, hybrid recommendation systems that utilize both LLMs and collaborative filtering algorithms provide more relevant recommendations.
The Future of Machine Learning Engineering
While generative AI currently dominates discussions, traditional machine learning retains its value. According to Gartner, 60% of companies will continue to use classic machine learning models alongside generative AI by 2025. The key is the seamless integration of both approaches to meet the specific needs of businesses.
Conclusion
Machine learning engineering, although challenged, is not obsolete. It is evolving and adapting to new technological realities. For tech decision-makers and entrepreneurs, understanding these dynamics is crucial to leveraging the opportunities offered by both worlds.
Let's discuss your project in 15 minutes.