Small Is the New Big 🚀

The industry often tells you that you need the largest, latest models. However, for many real workloads, this can result in unnecessary costs and latency for capabilities that you may not utilize. This session advocates for a different approach: using smaller, distilled models — finely tuned for your specific tasks and executed on your own infrastructure. These models can deliver quicker responses, lower costs, and keep your data secure within your account.

What to Expect 🛠️

• Self-hosting open-weight models on EKS.

Who It's For 💼

• Developers and technical founders preparing to run AI in production, focusing on cost and latency rather than just demos.

What You'll Leave With 📝

• A straightforward framework for determining model size per workload.
• A back-of-envelope cost model for your specific metrics.
• A clear roadmap to achieve faster, cheaper, and private inference.

This event is part of Basecamp, a community-powered tech week in Bengaluru aimed at bringing together founders, builders, and operators. Basecamp is free for the community and driven by community support.

Join the event here! — registration is required.