38M-param time-series world model: FSQ tokenizer → Mamba-2 JEPA → OT-CFM → TD-MPC2 agent. 838M tokens, TPU v6e, JAX/Flax.

NIFTY 50 5-day trend classification using Decision Tree, Random Forest and Logistic Regression with live prediction system.

Deep learning pipeline for financial time-series forecasting using LSTM, CNN, CNN–LSTM and ResNet–LSTM with Gramian Angular Difference Field (GADF) encoding and an interactive Streamlit dashboard.

End-to-end ML pipeline that predicts BTC/USDT price direction (4h horizon) using XGBoost + Optuna + SHAP. 9-phase architecture, Walk-Forward Validation across 15 folds, 37 technical indicators, 98 automated tests. ROC-AUC: 0.5431.

Graph + temporal deep learning for cross-sectional S&P 500 ranking. 9-variant ablation, 224 tests, val IC 0.0284 on yfinance data.

Reinforcement-learning aggregator on top of the TradingAgents multi-agent LLM trading framework. PPO policy that beats buy-and-hold on the 2026 YTD test across two LLM backbones (Anthropic + OpenAI); cross-LLM transfer holds. Course project, Columbia IEORE 4733.

Real-time financial sentiment pipeline: 3-model FinBERT ensemble (86.7% acc, neg PR-AUC=0.908) · MC Dropout uncertainty · TimescaleDB · Kafka · FastAPI · MLflow · Prefect 2

Advanced ML system combining LSTM attention networks, Transformer architectures, and gradient boosting ensembles for financial time series forecasting

Transformer‑based Bull/Bear classifier for Bitcoin using long‑window trend features and pretrained inference‑only weights.

Hesitation-weighted Transformer perplexity for unsupervised accounting anomaly detection — validated on SEC AAER, 8-K Item 4.02, and benchmarked against Beneish & Altman.

Coursera DS Coding Challenge — loan default prediction on 109K borrowers using PCA + Logistic Regression pipeline, evaluated on ROC AUC with competition-format submission

FinFusion: S&P 500 return forecasting with Temporal Fusion Transformers - compares TFT, ARIMAX, LSTM, and regime-aware variants.

Intelligent loan approval system using Support Vector Machine (SVM) for automated credit assessment and loan status prediction

ML pipeline for bank insolvency prediction (1–5 years). 4 ensemble models + SHAP explainability. Production-ready with Docker, reproducible research.

Trabajo de Fin de Grado en Ingeniería Matemática: Sistema de predicción direccional de Bitcoin mediante modelos de machine learning (LightGBM) y análisis de sentimiento (RoBERTa). Investigación sobre integración multimodal en mercados financieros.

Deep RL agent for financial market signal generation — PPO/A2C/SAC/TD3, 99 indicators, ensemble signals, 4-level quality gate

Advanced gold price forecasting system beating academic benchmarks with 9+ ML models. Features rolling window predictions, real-time analytics dashboard, and extensible architecture. Built with uv, FastAPI, and Next.js for cross-platform performance.

Bitcoin trading agent using Deep Q-Learning and synthetic market scenarios.

Real time fraud detection pipeline

Production-grade ML signal intelligence engine for quantitative trading. Powers real-time XGBoost inference across 100 S&P 500 tickers, 4-agent decision governance, algorithmic drift detection with automatic exposure scaling, and geopolitical risk overlay via live news APIs.