In our introductory post, “From ML’Oops’ to MLOps for Energy”, we highlighted three key
problems for effective Data Science in Energy. They are:
- Citizen Development
- High-Performance Computing
- Federated Connectivity for Generative AI
Having outlined how deep the rabbit hole goes, let’s now climb ourselves back out by analyzing
each of these topics in more detail. Again, we at Activera Consulting have worked directly with
one of our ecosystem partners, Domino Data Lab, to focus on Energy-specific challenges and
provide some insight into how to solve for them.
We were recently talking to a large Fintech company and they were talking about their
Generative AI efforts. Initial passes were taking place through OpenAI to prove out use-cases
but the goal of the program was to then go Open Source to recreate using proprietary data,
improve efficacy, and reduce costs of the models.
This is a trend consistent with other conversations we’ve had with Energy firms. Fine-tuning
Large Language Models (LLMs) while retaining ownership, model control, and data security are
key considerations for the savvy. But beyond just the build/buy conversation, which we will
address in the conclusion, there are multiple focus-items on the minds of those progressing Gen
AI in their organizations.
