Bite-sized steps for each task with exact file paths, code snippets,
verification commands, and parallel execution strategy.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Safety, live trading, security, and observability gaps identified
from comprehensive codebase audit. Organized for parallel swarm execution.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
8-task TDD plan to bridge 8 disconnected layers in the backtesting
pipeline: DBN converter, replay engine, 51-dim feature wiring,
model loader, registry startup, position tracking, PnL tracking,
and end-to-end integration test.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
13-task plan covering data pipeline fixes (Databento stream, data acquisition,
DBN uploader) and trading service core wiring (RiskEngine, MLEngine, VaR,
feature extraction, order matching, API gateway, Prometheus panics).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
8-task plan to bridge disconnected backtesting pipeline layers:
DBN parser → feature extraction → ML inference → position tracking → PnL
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Covers data pipeline, trading service core, broker connectivity
(FIX 4.4 + TWS), execution algorithms, security/compliance, and
observability — based on comprehensive non-ML audit.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Two-phase design:
A) Real 6E.FUT data → 4-model ensemble → PaperBroker → P&L metrics
B) Best hyperopt params → train → deploy to ensemble → ValidationHarness
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Create PpoStrategy and PpoLstmStrategy implementing ValidatableStrategy
to run PPO through walk-forward validation with DSR, PBO, and permutation
tests. Both variants validated on real 6E.FUT data (29,937 bars, 15 folds).
Key implementation detail: LSTM hidden states are detached from the
computation graph after each step to prevent stack overflow from
unbounded graph growth across 30k+ sequential forward passes.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Single integration test that verifies the complete train → checkpoint →
validate pipeline works on real 6E.FUT data. 7 assertions covering
loss convergence, Q-value divergence, checkpoint round-trip, epsilon
decay, and Sharpe improvement vs untrained baseline.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Research-validated design for fixing critical DQN issues:
- CQL regularization for offline RL training on historical data
- IQN integration replacing broken C51 (Candle scatter_add bug)
- CVaR risk-aware action selection
- State dimension consolidation (51 is canonical)
- NaN panic fix, Adam epsilon fix per Rainbow paper
Verified against 20+ papers and 2026 SOTA.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
16-task bottom-up execution plan with exact file paths, commands,
and safety gates. Covers: doc purge, dead code removal, module
consolidation, test reorganization, file splitting, and folder
restructuring.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Comprehensive cleanup plan covering: documentation purge, dead code
removal (~4,846 lines across 11 files), module consolidation,
test reorganization, large file splitting, and folder restructuring.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
