Community: setup deepseek on aws for a private model coding assistant

What I really want to do now is have unlimited coding AI assistance and I am contemplating setting up DeepSeek V3 on an EC2 instance with GPUs (g5 or g6 instances) and wrapping it with an API that gives me unlimited access to the model via a coding assistant like Cline. This might cost me $300 a month at AWS but it’s my own and unlimited without rate limits and my data going to companies.