Skip to content
View datascientistandAI's full-sized avatar
2 followers · 1 following

Block or report datascientistandAI

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.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

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

Report abuse
datascientistandAI/README.md

Hi, I'm Boris Garcial 👋

Computer scientist focused on machine learning systems at the intersection of AI and financial markets — forecasting models, trading decision architectures, and the data infrastructure that feeds them.

What I work on

  • Deep learning for time-series forecasting (sequence encoders, quantile/distributional prediction)
  • Reinforcement learning and decision-making systems for trading
  • Market regime modeling and mechanical (rule-based) decision layers built on top of learned forecasts
  • Data engineering: live multi-source ingestion pipelines with point-in-time correctness and automated validation gates

Featured project

film-bpc-macro-trader — a world-model architecture for FX trading built around quantile rollout prediction and a mechanical, auditable barrier-simulation layer, validated across 279 walk-forward folds on 2019–2026 data. A research/engineering showcase; core logic and weights are kept private as the system is used in active trading.

Currently learning: SQL for database design/querying, web data collection, and Tableau for dashboarding.

Stack: Python, PyTorch, scikit-learn, LightGBM, DuckDB, C++

Pinned Loading

  1. continual-learning-controller continual-learning-controller Public

    Gradient resurrection: a theoretically-derived, trust-scaled controller for continual learning, with full proof and an honestly-reported failed extension hypothesis

    Python 1

  2. r-tokenizer-regime-validation r-tokenizer-regime-validation Public

    A learned market-state representation validated against two regime-detection baselines on synthetic positive-control data. Description + results only — no code.

    1