679a483de5
Task 4 — CVaR Action Selection: - Add use_cvar_action_selection and cvar_alpha fields to DQNHyperparameters - Wire from hyperparams into DQNConfig constructor (was hardcoded to false) - Default: enabled (true) with alpha=0.05 (worst 5% quantile tail) - Unblocks risk-aware position scaling via IQN head's compute_cvar_q() Task 5 — Curiosity Wiring (verified active): - GpuCuriosityTrainer trains forward model on GPU experience data - train_curiosity_gpu() called from training_loop after experience collection - curiosity_weight=0.05 (Task 2) gates trainer creation — active when >0 - Intrinsic reward injection into DQN kernel deferred (Phase 4+, per kernel docs) Task 6 — CQL Conservative Loss GPU Kernel: - Add use_cql/cql_alpha to GpuDqnTrainConfig (wired from DQNHyperparameters) - Implement cql_logit_grad_kernel: computes dCQL/d_logits for Branching Dueling C51 - Per-branch logsumexp penalty with softmax gradient through expectation chain - One thread per sample, iterates 3 branches (exposure, order, urgency) - Add apply_cql_gradient() method: launches CQL kernel + cuBLAS backward_full - Accumulates CQL parameter gradients into grad_buf (beta=1.0) - Same injection pattern as IQN trunk gradient - Wire into FusedTrainingCtx::run_full_step() between graph_forward and graph_adam Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Foxhunt
Production HFT trading system in Rust.
Architecture
The workspace contains 32 crates organized as follows:
Core Libraries (16)
| Crate | Purpose |
|---|---|
trading_engine |
Order processing, FIX 4.4, IB TWS, SIMD, RDTSC timing |
risk |
VaR, Kelly, circuit breakers, kill switches, compliance |
risk-data |
Risk data types and shared structures |
trading-data |
Trading data types |
ml |
DQN Rainbow, PPO, TFT, Mamba2, ensemble inference |
ml-data |
ML data types and feature definitions |
data |
Market data ingestion and storage |
backtesting |
Replay engine, strategy tester |
adaptive-strategy |
Ensemble execution, microstructure analysis |
common |
Shared types, resilience, error handling |
storage |
S3 and local model storage |
model_loader |
Model serialization and loading |
market-data |
Market data feed handlers |
database |
PostgreSQL access layer (SQLx) |
config |
Configuration management |
tli |
CLI commands and tooling |
Services (8)
| Service | Purpose |
|---|---|
backtesting_service |
gRPC backtesting service |
broker_gateway_service |
FIX routing, broker connectivity |
trading_service |
Core trading operations |
ml_training_service |
Model training orchestration |
data_acquisition_service |
Market data acquisition |
trading_agent_service |
Autonomous trading agents |
api_gateway |
gRPC API gateway with auth |
web-gateway |
Axum REST + WebSocket gateway |
Frontend
web-dashboard/ -- React 19 + TypeScript + Vite + TradingView charts.
Building
# Check compilation (no PostgreSQL required)
SQLX_OFFLINE=true cargo check --workspace
# Run tests for a specific crate
SQLX_OFFLINE=true cargo test -p <crate> --lib
# Clippy
SQLX_OFFLINE=true cargo clippy --workspace
ML Models
Four production model architectures on Candle v0.9.1 with CUDA:
- DQN Rainbow -- Deep Q-Network with prioritized replay, dueling heads, noisy nets
- PPO -- Proximal Policy Optimization with GAE and LSTM policies
- TFT -- Temporal Fusion Transformer for multi-horizon forecasting
- Mamba2 -- State space model for sequence prediction
Each model has a standalone trainer and a UnifiedTrainable adapter for the hyperopt pipeline.
Infrastructure
- Git: Gitea at
git.fxhnt.ai(Tailscale-only), Scaleway DEV1-S - Observability: OpenTelemetry OTLP (env
OTEL_EXPORTER_OTLP_ENDPOINT) - Database: PostgreSQL with SQLx offline mode for CI
License
Proprietary. All rights reserved.