Erik An

BSc Mathematical Foundations of Data Science.

Data science specialization.

Software engineering. Completed in 10 months, first team to finish in a cohort of 2024.

A-Levels in Mathematics, Physics, and History.

A version-controlled, executable knowledge base for a mathematics degree — lecture notes, proofs, and problem sets in Org-mode and LaTeX with embedded Julia, Python, and C++. Built a prompt specification system with defined input/output contracts. As of May 2026, ~260k words, 18,000+ LaTeX fragments.

A Shazam-style audio identification engine — spectrogram analysis, spectral peak extraction, and combinatorial hashing matched against a SQLite fingerprint database. 97.8% identification on 138 noisy clips against an 8,000-song database, ~260 ms per query.

A real-time multiplayer Pong platform — microservice backend (Fastify/SQLite), TypeScript frontend, remote multiplayer, and a single-player AI opponent with constrained perception. Team project, deployed via Docker.

A non-blocking HTTP server from scratch in C++98 — GET/POST/DELETE, CGI execution, file uploads, and NGINX-style config parsing, built on a single epoll event loop. Incremental state-machine request parser; cookie-based sessions. Team project.

The JPEG compression pipeline implemented from the raw mathematics — manual 1D DCT summation extended to a separable 2D transform, block processing, quantization, and inverse reconstruction. Self-directed, to internalize the DCT covered in Linear Algebra.

Univariate linear regression from scratch — gradient descent, feature normalization, and model evaluation, with a Makefile-driven train/predict/evaluate pipeline.