I'm a Data Scientist and ML Engineer who bridges the gap between high-performance systems programming and modern AI architectures. I specialize in building machine learning algorithms from the ground up, designing automated ETL pipelines, and developing LLM-powered data applications.
Currently studying at Davidson College, I'm passionate about the intersection of Quantitative Economics, System Design, and Data Engineering.
- Languages: Python, C++, SQL, C, R, Java, Bash/Shell
- AI & Machine Learning: Scikit-Learn, Deep Learning (MLPs), RAG, Google Gemini, XGBoost, Pandas, NumPy
- Data Engineering & DevOps: PostgreSQL, pgvector (Vector DBs), Docker, GitHub Actions (CI/CD), ETL Pipelines
- Build Tools & Environments: Linux Command Line, CMake, Make, Git, pybind11, Streamlit
- Tech Stack:
C++17,Python,pybind11,CMake,AVX2 SIMD - Built a complete ML framework entirely from scratch in C++ (no PyTorch, no Eigen). Implemented forward/backward passes, Multi-Layer Perceptrons (MLPs), and the Adam optimizer.
- Accelerated matrix operations utilizing explicit hardware-level AVX2 SIMD intrinsics and exposed the high-performance backend to Python via custom bindings.
- Tech Stack:
Python,Google Gemini,Supabase (PostgreSQL/pgvector),Docker,Streamlit - Engineered an automated ETL pipeline that aggregates daily macroeconomic data and news, using an LLM to synthesize a professional portfolio manager brief.
- Architected a semantic search chatbot utilizing 768-dimensional text embeddings, completely containerized with Docker and scheduled via GitHub Actions.
- Tech Stack:
Python,Streamlit,Economics - Developed an interactive dashboard to visualize and simulate shifts in the IS-LM macroeconomic model.
- Developing market regime-switching detection algorithms using Hidden Markov Models.
- Deploying Autonomous Dockerized Openclaw Agents through VPS.
- Deepening my knowledge of low-level system design and large-scale data lake architecture.
- LinkedIn: Alexander Shields
- Email: alshields1@davidson.edu