Looking for people interested in AI × Infrastructure × Sustainability research
Looking for people interested in AI × Infrastructure × Sustainability research
I've been exploring an area that I find increasingly interesting: the environmental and infrastructure impact of AI, and what we could potentially do about it.
AI isn't just a software problem anymore. Behind every model are GPUs, data centers, electricity, cooling, water, networking, hardware manufacturing, and increasingly complex infrastructure.
I'm interested in exploring questions such as:
- How do we accurately measure the energy, water and carbon footprint of AI workloads?
- Can AI inference be made significantly more resource-efficient?
- Could workloads be scheduled based on energy availability, carbon intensity or water stress?
- Do we actually need large models for every task?
- Can smaller models, routing, caching, quantization, batching, etc. meaningfully reduce resource consumption?
- Can AI itself help optimize energy grids, water systems, agriculture, conservation and other environmental problems?
- What are the gaps in current research?
- What interesting research questions haven't been explored enough yet?
At this stage, I'm not trying to sell a specific idea or claim to have the answers.
I'd like to find people who are interested in:
• Reading and analyzing research papers
• Discussing existing findings and identifying gaps
• Challenging assumptions and ideas
• Brainstorming potential research questions
• Designing small experiments or prototypes
• Potentially collaborating on a research paper/project in the future
You don't necessarily need to be an academic researcher. I'm interested in people from AI/ML, backend/infrastructure, distributed systems, data centers, energy, sustainability, environmental science, or related fields.
I'm approaching this as an open exploration: research → discussion → experiments → ideas → potentially a paper/project.
If this sounds interesting to you, feel free to comment or DM me. I'd be happy to start with a small group and see where the discussions lead.
Would love to hear what problems you think are currently being overlooked in AI's environmental/infrastructure impact.