https://deepnet.one Deepnet One is an applied AI research and development company. Its R&D lab works on neural-based reasoning (the Engram programme); it also maintains the Relearn reinforcement learning library, builds the Assistatron trading assistant and develops small, fast AI models for algorithmic trading. en https://deepnet.one/post/seventeen-years-of-algotrading/ https://deepnet.one/post/seventeen-years-of-algotrading/ Tue, 03 Nov 2026 00:00:00 GMT Twenty-six years of watching the market, seventeen of trying to beat it, the last few with AI. This is the record so far: what I built, what the market thought of it and what is running today. https://deepnet.one/post/point-in-time-text-classification/ https://deepnet.one/post/point-in-time-text-classification/ Wed, 24 Jun 2026 00:00:00 GMT A forward-looking prediction problem that is really a timeseries: point-in-time profiles, an ever-growing text space, and why the hard part was leakage and out-of-sample validation, not the model. https://deepnet.one/post/tiny-bert-in-production/ https://deepnet.one/post/tiny-bert-in-production/ Mon, 18 May 2026 00:00:00 GMT Four independent classifiers converged on the same recipe: a 2-to-4 layer BERT quantized with OpenVINO, running in single-digit milliseconds on a CPU with no GPU. How, and why the data split mattered more than the model. https://deepnet.one/post/when-the-simpler-model-won/ https://deepnet.one/post/when-the-simpler-model-won/ Thu, 30 Apr 2026 00:00:00 GMT What a failed financial VAE taught us about representation learning. Across four model families and six trained models, the simplest model reconstructed best, and every layer of added sophistication made it worse.