I'm a Computer Science student and software engineer (MSc CS @ TU Delft, BSc CS @ Maastricht University). I previously worked as an AI Developer at Oraspire, and I build reproducible, math-heavy systems across Machine Learning, Operations Research, Quantitative Finance, and Systems Engineering.
- Industrial-grade optimizer: a vehicle routing optimizer with MILP formulations, Clarke-Wright and 2-opt/Or-opt heuristics, and a seeded benchmark harness that reports solution quality against runtime.
- Maritime Tech Radar: a config-driven, human-reviewed technology radar for port and maritime start-ups, with a logged collect-extract-review-score pipeline and a Streamlit dashboard.
- BSc Thesis, Streaming Signal Decomposition: a Python framework that decomposes noisy real-time signals window by window (SSD-style, windowed embedding) under tight memory/latency constraints, with a YAML-driven benchmark pipeline and stability metrics that track whether components stay consistent as the signal changes.
- Limit Order Book simulator: a market-microstructure toy model comparing market and limit orders across spread and impact regimes, with a game-theoretic layer on liquidity provision versus liquidity taking.
- Operations Research (NTU specialization): LP/IP, duality, network flows, Lagrangian relaxation.
- Recently completed: Game Theory (Stanford), plus Micro, Macro and Business Economics.
- Up next: Financial Engineering and Risk Management, systems engineering (MBSE), and GCP MLOps.
- Healthchain Access Control: a decentralized, consent-driven access-control stack (Solidity/Hardhat) for patient healthcare data, with large-scale automated simulation tooling.
- Transitor Routing Engine: a Java-based public transport routing engine integrating GTFS datasets to compute optimal transit routes and accessibility metrics.
- Modular AI Tutor: a real-time hint-generation system for robotics students, with GDPR-compliant architecture and performance analytics.
- E-Voting Systems: a comparison of a client-server design and two blockchain-based architectures on reliability, scalability and voter privacy.
- Languages: Python, Java, C, SQL/NoSQL, Solidity, JavaScript/TypeScript
- ML & Data Science: PyTorch, TensorFlow, Pandas, NumPy, Scikit-learn
- Optimization & Apps: MILP modelling (PuLP/CBC), Streamlit, Plotly
- Systems & Tools: Git, Linux, Hardhat, Express, JUnit, Pytest, uv
- Concepts: Time-Series Analysis, Signal Decomposition, Applied Optimization, Market Microstructure, Smart Contracts, Reproducible Benchmarking
- LinkedIn: linkedin.com/in/sachaloeb
- Portfolio: check out my pinned repositories below for code samples, benchmark reports, and system architectures.