Our AI models
Proprietary models
We developed over a dozen proprietary ML models, which are trained on 240TB datasets, live order books of exchanges and ETH network data. For over a year we trained our AI LLM & RL on cutting-edge servers each loaded with 10xNvidia A40 GPUs.
With our own self-trained LLM models, AI is not just a buzzword for us, we have skin the game as we have self-funded the model training and ML orchestration for many months.
We don’t use any third parties for our setup. We source the data ourself with lowest latency possible, we handle all trades with our own proprietary trading engine and sufficient failover safety. To have a setup that we entirely control ourself and know inside-out is necessary to set new standards in this sector.
Last updated