FAIQ HAMMAM MUTAQIN
Yogyakarta, Indonesia cyclocerine.me orcid.org/0009-0008-4920-7623 github.com/cyclocerine linkedin.com/in/faiqhammam hammamfaiq@protonmail.comProfessional Summary
Performance-driven AI/ML and Full-Stack Engineer with hands-on experience in PyTorch, CUDA C++, React, Django REST Framework, and Docker. Builds quantitative trading models, PatchTST forecasting pipelines, PPO agents, and automated backtesting systems. Experienced in CUDA kernel optimization, REST API development, and containerized web applications.
Education
Universitas Pembangunan Nasional "Veteran" Yogyakarta (UPNYK)
Yogyakarta, IndonesiaBachelor of Science in Informatics · Expected Graduation: 2028
- Current GPA: 3.44 / 4.00
- Relevant Coursework: Machine Learning, Deep Learning, Data Structures & Algorithms, Web Engineering, Database Management Systems, Operating Systems, Parallel Programming.
Technical Skills
- AI/ML & Quantitative Modeling: PyTorch, CUDA, Deep Reinforcement Learning (PPO), PatchTST, Scikit-Learn, Pandas, NumPy, Time-Series Forecasting, Automated Backtesting.
- Web Technologies: Next.js, Vite, React, Django, Django REST Framework, HTML5, CSS3, Tailwind CSS, REST API Development.
- Programming Languages: Python, C++, SQL, JavaScript, TypeScript, C.
- Tools & Environment: Git, GitHub, Docker, Linux (Debian/Ubuntu), CUDA Toolkit.
Featured Projects
Q-ALU PyTorch Implementation
PyTorch, CUDA, Python- Engineered and open-sourced a PyTorch implementation of Q-ALU (Quant Asymmetric Leaky Unit) to model downside-risk protection in high-frequency quantitative trading workflows.
- Evaluated PPO trading policies on 5,000 out-of-sample ticks with 1.0–1.5 bps turnover fees; Q-ALU achieved 75.85% return, 0.142 per-step Sortino, 2.09 profit factor, and 2.79% maximum drawdown in the reported experiment.
- Tested across 2,834 BTC/USDT market bars and five random seeds; implemented CUDA C++ kernels with a measured 0.308 ms backward execution time.
AI Algorithmic Hedge Fund Platform
PatchTST, PPO, Python- Built a modular algorithmic trading system combining PatchTST forecasting, Bayesian hyperparameter tuning, and PPO-based trading decisions across stocks, crypto, forex, and commodities.
- Engineered a unified prediction and backtesting pipeline with 14+ technical signals, transaction-cost and slippage modeling, risk management, and automated Sharpe, drawdown, win-rate, and return metrics.
- Reported BMRI.JK backtest results (PatchTST tuned, PPO trained for 200 episodes) of +240% portfolio return; this figure reflects the project's documented backtest, not live trading performance.
Full-Stack Interactive Web Platform
Next.js, Vite, React, Django- Built and deployed a responsive full-stack application with Next.js, Vite, and React, using modular frontend components.
- Developed Django REST Framework endpoints and database models, with secure authentication for API access.
- Delivered mobile- and desktop-responsive interfaces using Tailwind CSS.
Stock Price & Pattern Prediction System
Python, Scikit-Learn, ML- Developed a Python and Pandas machine-learning pipeline for candlestick-pattern recognition and financial time-series analysis.
- Engineered dynamic technical indicators and custom neural-network features to identify price-reversal patterns across financial instruments.
- Applied feature engineering to reduce market noise and support classification performance.
Organizations & Activities
Information Technology Student Community / Peer Tutor
Yogyakarta, IndonesiaUPN "Veteran" Yogyakarta · 2023 – Present
- Mentored junior engineering students in Data Structures & Algorithms, C++, Python programming, and Web Development to strengthen core technical fundamentals.