Research Workflows for Carbon-Optimal AI Inference Scheduling
For this project, I applied the ChatGPT workflows from the masterclass to streamline the research phases for a system designed to optimize AI inference scheduling.
-
Brainstorming & Ideation: Used structured prompting to generate foundational concepts for a grid-aware LLM energy agent.
-
Literature Review: Summarized dense academic papers regarding machine learning and predictive analytics to quickly identify gaps in current carbon-reducing methodologies.
-
Outline Generation: Created a comprehensive framework and structural foundation for the initial proposal.
Implementing these workflows drastically reduced the time spent synthesizing preliminary research and provided clear direction for the project's development.