Finetuning to Inference: Serve Open Models with Tensorfuse & AWS

​Join Tensorfuse and AWS at the GenAI Loft in San Francisco for a hands-on workshop on fine-tuning and deploying open source LLMs like Llama, Qwen, Deepseek, in your AWS VPC.

​Unlike other events, this will be a co-build session where you will:

• ​Use Tensorfuse to provision an EKS cluster in your AWS VPC optimised for running AI workloads.

• ​Customise open source LLMs by fine-tuning LoRA adapters

• ​Deploy the fine-tuned model and run inference using vLLM

​Best Part: For this workshop, we will be using readily available on-demand GPUs like A10G and L40S on AWS.

​Agenda

• ​2:00-2:30 p.m. | Check-in and brief about Tensorfuse

• ​2:30-4:30 p.m. | Deploy-athon: We will configure the Tensorfuse runtime in your AWS account and fine-tune and deploy Llama models

• ​4:30-5:00 p.m | Networking, Snacks and Drinks!

​Pre-requisite

​Everyone interested in customizing and self-hosting LLMs is welcome. However, you’ll get the most from this event if you:

• ​Have an active AWS account

• ​Have sufficient on-demand GPU quotas for G and VT instances (ideally in us-east-1 or us-west-2). Check your quotas here

• ​Permissions to configure resources in your AWS.

​We have limited spots (~100), book your seat now.

Excited to host you all at the GenAI Loft in SF!