Adrian’s AI/ML 2021 Highlights
I think 2021 has been a wonderful year in terms of “AI maturity” around the world. Here are my main AI/ML 2021 highlights (not a universal truth, so please feel free to disagree and/or add other points) ⬇️ ⬇️
- Democratization of the AI adoption is generating new opportunities for all kinds of AI professionals (tech, biz, and hybrid profiles). I see more people with “atypical profiles” re/up-skilling so they can help with new internal AI initiatives for small and medium businesses. We are going more granular in terms of AI skills and responsibilities. Cloud-enabled / aaS platforms are helping a lot.
- MLOps is the key industry trend, and an organic answer to the need for a well-defined end-to-end AI/ML lifecycle. This is also contributing to clarify tasks / micro-tasks and responsibilities. I think data and ML engineers will have a lot of fun in 2022 :) their roles are evolving very quickly, same as their importance within the adopter organizations.
- Responsible AI is finally becoming something tangible. Main concern for AI professionals was how to actually evaluate responsible AI factors during the analysis, implementation, and prod phases. Big companies are releasing both commercial and open source solutions to check fairness, explicability, data bias, etc. Still a WIP but I really like what I see from Microsoft, TensorFlow/Google, etc.
- The data scientist skillset is also evolving. Some of the DS/DE tasks are becoming a bit more of a “commodity” (e.g., initial naive approaches, model benchmarking), and technical complexity requires more and better soft skills than ever. I have met wonderful data scientist, but my favourite ones are those who are able to explain choices and results in a very simple way, so everyone can understand the key pieces. Superstar data scientists will bring that to the companies, and they will become key decision influencers.
Undoubtedly, 2022 will be great for AI/ML professionals and adopter companies. What do you think?