DWQA QuestionsCategory: Artificial IntelligenceWhat are the hidden costs of building an algorithm team?
Titanium Engineer asked 1 month ago

Years of working in Silicon Valley have taught us that building an algorithm team is far more “expensive” than cognitive. For example, software engineers and scientists need to be equipped with software engineers to support them with tools, expensive GPUs and training machines such as DGX or Titan P level need to be purchased, and tagging teams and relevant project managers may also be required. Want to hear what AI practitioners think about this? What are the hidden costs of your company?

G_Koala_C replied 1 month ago

It’s expensive!

1 Answers
hold_baby answered 1 month ago

Remove AI team
The cost of a tagging system and its associated tagging staff, developers and product managers is not small
With the annotation system, you need material annotation. This is a headache. You need a lot of real data to meet your training task. So how to get the material cost is also considered, but it also varies from person to person. For example, you have something to do with ZF
With the above, your AI team will be able to work. According to the materials, different models will come out. There will be machines, bandwidth and time cost
When your AI team has output, you need a sales team to sell your model, and the cost of the sales team should not be underestimated
Customers have to show their own model style to the customer interaction project, but in the early stage, you will find that the accuracy of the model is a bit harsh, so at this time, you need to have an audit platform to audit which data can be shown to customers and which can’t be shown to customers. After all, artificial + Intelligent = artificial intelligence
The cost of audit system is audit team + development + product + time
That’s the cost of our company right now