Model Catalog

Institutional-grade neural networks packaged for integration. Below: specs, deliverables, and what you get in a deployment bundle.

Price Prediction — LSTM

Directional multi-horizon forecasting

  • Type: LSTM ensemble (sequence-to-sequence)
  • Inputs: OHLCV, engineered features, exogenous signals
  • Outputs: Probability of direction at multiple horizons
  • Runtime: Docker image with TensorFlow 2.x; CPU & GPU variants
  • Deliverables: Model weights, Dockerfile, sample ingestion pipeline, evaluation notebook
  • Support: 2-week onboarding + optional SLAs
Validated Sharpe-like metric: confidential Request Quote

Volume Analysis — CNN

Pattern recognition on volume & microstructure

  • Type: 1D/2D convolutional networks
  • Inputs: tick/volume bars, time-of-day features
  • Outputs: event classification and anomaly flags
  • Runtime: Lightweight inference optimized for streaming
  • Deliverables: exportable ONNX model, integration guide
  • Support: Integration workshop option
Convolutional Neural Network CNN Request Quote

Signal Classification Transformer

Context-aware buy/sell signal generator

  • Type: Transformer encoder with calibrated outputs
  • Inputs: feature sequences, events
  • Outputs: classification + confidence
  • Runtime: Docker + example microservice
  • Deliverables: model, validation suite, calibration artifacts
  • Support: customization packages available
Calibration: Platt-scaling included Request Quote

Custom Development

Bespoke models tailored to your alpha

  • Process: Discovery → prototype → validation → production
  • Deliverables: codebase, docs, reproducible pipelines
  • Engagement: fixed-bid or scoped retainer
Start a Conversation

Get technical docs & deployment packet

Or email tensorcog@gmail.com