Skip to content
View spirosrap's full-sized avatar
πŸ€–
πŸ€–

Block or report spirosrap

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
spirosrap/README.md

Spiros Raptis

Systematic Trader Β· AI Systems Builder

Engineering transparent, versioned trading systems with circuit breakers, regime logic, and psychological integrity.

πŸ“ Thessaloniki, Greece Β· X @srdevb

Profile views GitHub followers Collaboration


What I’m Building Now

  • πŸ“ˆ Versioned Bitcoin execution systems (v1.2.1f β†’ scaling ladders with circuit breakers).
  • 🧠 LLM-driven diagnostics for MAE/MFE integrity, expectancy, and emotional overrides.
  • πŸ§ͺ Stability under pressure: controlled exposure, freeze windows, post-trade review protocols.
  • πŸ”§ Clean research-to-production flow: notebooks β†’ Linux scripts β†’ tmux pipelines β†’ watchdogs.

Signature Engineering

πŸ”Ή Automated Trading Stack β€” v1.2.1f

A disciplined BTC execution engine featuring:

  • RSI signal with 1-bar confirmation
  • Regime detection (ATR% + Trend Slope)
  • Three pre-entry rejection filters for weak/uncertain regimes
  • Live MAE/MFE watchdog
  • Circuit breakers calibrated to emotional capacity
  • Structured scaling protocol (Step 1 β†’ Step 2 β†’ monitored checkpoints)

All changes occur only at predefined checkpoints. No mid-cycle optimization.


Projects

Estimation, control, and planning fundamentals. Foundation for disciplined engineering.

Research exercises and early explorations in feature engineering and ML-driven signals.

PyTorch implementations of RL algorithms used to build intuition for dynamic decision systems.

CLI tool for tracking review queue positions with clean, stable utility code.


Toolbox

Python pandas NumPy PyTorch
scikit-learn Jupyter
TradingView Pine Script Coinbase Perps
Linux Docker tmux
VS Code Cursor


Operating Principles

  • πŸ›‘οΈ Risk first: position sizing, SL/TP asymmetry, circuit breakers.
  • ❄️ Frozen rules: modify only at checkpoints with sufficient data.
  • πŸ“ Track edge: expectancy, MAE/MFE, regime shifts, volatility context.
  • 🧠 Build simple systems you fully understand before scaling complexity.
  • πŸ”¬ Treat losses as data, not identity.

Current Focus

  • πŸ€– LLM-assisted journaling, diagnostics, and signal triage.
  • πŸ“‘ Automated post-trade evaluation with regime tagging and edge tracking.
  • 🧩 Meta-system design: versioning, freeze windows, scaling ladders.
  • πŸ”’ Reinforcing psychological stability under increased margin conditions.

Signals

GitHub Stats Top Languages

Where disciplined execution meets emotional resilience.
If you want to build something measurable and real, reach out.

Pinned Loading

  1. flyingcar flyingcar Public

    Flying Car Nanodegree

    C 42 27

  2. Deep-Reinforcement-Learning Deep-Reinforcement-Learning Public

    Deep Reinforcement Learning - Implementations and Theory: A path to mastery

    Python 12 3

  3. AI-for-Trading AI-for-Trading Public

    Artificial Intelligence for Trading Nanodegree

    HTML 11 9

  4. Artificial-Intelligence-Nanodegree Artificial-Intelligence-Nanodegree Public

    Artificial Intelligence portfolio

    Jupyter Notebook 2