Daniel G. Alfert
Copenhagen · danielgalfert@gmail.comProfile
Recent graduate based in Copenhagen with dual BSc degrees in Mathematics and Computer Science and an MSc in Quantum Information Science. My strengths are at the intersection of quantitative thinking and engineering: statistical modeling, clear reasoning under uncertainty, and shipping working software (APIs, authentication, deployment).
Looking for roles in AI engineering, model validation, quantitative development, or data and AI consulting, where learning fast and working rigorously are both part of the job.
Experience
- Built a greenfield internal knowledge platform for researchers (2-person team, minimal guidance, ~12 h/week).
- Implemented ~80% of the codebase across frontend and backend; focused on reliability and clarity over cleverness.
- Designed authentication flows using JWT access + refresh tokens with secure HTTP-only cookies.
- Developed async APIs with FastAPI and managed relational data (~15 tables) using SQLite.
- Delivered a working prototype deployed on an internal machine for ~50 internal users.
- Participated in an internal applied AI program focused on end-to-end delivery rather than isolated modeling.
- Trained and deployed a computer vision classification model via Azure ML to edge devices using Azure agents.
- Gained first-hand experience of deployment constraints: foreign machines, agents, and operational friction.
- Studied how noise impacts Gibbs states prepared by shallow local quantum circuits (k-local, depth O(log n)).
- Implemented tensor-network simulations in Python using Quimb (systems up to 8 qubits).
- Focused on depolarizing noise (with discussion of amplitude damping) and formulated an original conjecture.
Projects
- End-to-end retrieval pipeline: CV ingestion, structured LLM extraction, semantic indexing in ChromaDB, hybrid retrieval via reciprocal rank fusion, and an explainability layer.
- Delivered as a live panel demo showing how retrieval and explainability support staffing decisions.
- Implemented classical time series models and explored stepwise model selection.
- Ran diagnostics including residual checks and heteroskedasticity tests.
- Regression: linear models with L1/L2 regularization, hyperparameter tuning via grid search.
- Unsupervised learning: Gaussian Mixture Models on physiological signals to explore relationships with emotional states.
- Self-hosted services on a Raspberry Pi using Linux, Docker, and systemd.
- Nginx reverse proxy and HTTPS via Certbot; Tailscale for secure remote access.
- Debugged real operational issues (e.g., DNS failures when Pi-hole went offline).
Education
- Relevant focus: time series analysis, statistical modeling, computational methods, research-grade simulation.
- Thesis: learning local Gibbs states from noisy shallow quantum circuits (tensor networks, Quimb).
- Coursework in probability, statistical inference, linear algebra, algorithms, and stochastic processes.
- BSc thesis: fundamentals of quantum computing; small algorithm simulations in Qiskit.
Skills
Python (NumPy, Pandas, scikit-learn, Quimb, Qiskit), C++ (STL; algorithmic practice), JavaScript (Vue). SQL basics + ORMs.
Time series (AR/ARIMA/ARIMAX), regression (L1/L2), Gaussian Mixture Models, hypothesis testing, Monte Carlo (European option pricing coursework).
Linux, Git/GitHub, GitHub Actions, Docker, Nginx, systemd, TLS/Certbot, reverse proxying, LLM APIs, vector databases.
Prioritize correctness and maintainability, write things down, make failures loud rather than silent. Comfortable in ambiguous environments.