- Challenge
- Training a neural network burns GPU time before you know whether the network is worth it.
- Action
- I co-founded Nixtee and I run it as CEO. The platform predicts network performance before training and saves compute.
- What changed
- Member of NVIDIA Inception.
- What it proves
- A decision can come before the compute. Not after the capacity is burned.
market
Nixtee
Market: network performance before the GPU burns. CEO, co-founder. NVIDIA Inception.
I co-founded Nixtee and I run it as CEO. The platform predicts neural-network performance before training.
Training burns capacity before you know whether the network is worth it. This is a decision before the compute starts.