Open Models, Sovereign AI, Systems, Scale and Product Research and Deployment Engineering
Work with open weights end to end: adapt, shrink and serve models, study sovereign AI, and carry systems from research into products and deployment.
- For
- Builders and engineers who learn by making things
- You will be able to
- Build working applications with small and open-weight models
- Adapt a model to a task — prompting, fine-tuning with LoRA, and distillation
- Make models small and fast — quantization, efficient inference and serving on modest hardware
- Reason about sovereign AI — open weights, national capability, and who depends on whom
- Evaluate honestly what a small model can and cannot do, with users in the loop
- Carry a system from experiment into a product and a deployment
- Keep a portfolio of models, Spaces, datasets and repositories that shows the craft
Hands-on craft, engineering and research with models that already exist. Every build pairs one module with one domain: healthcare and life sciences, biology, physical AI, scientific discovery, or automated research.
Questions about this course, or a case that argues with it? Write to mepc3197@columbia.edu · the title goes in the subject line
Related
- Complex Systems, Not Integrations: What Forward-Deployed/AI Engineering Is Missingessayscience · engineering
- Daybook, 1 September 2026dayengineering
- Our first job was forward deployed engineering — or maybe beyond thatessayengineering · philosophy