From College Project to 400 GitHub Stars: The Story of AIJack (medium.com)

AIJack is an open-source Python toolkit for experimenting with security and privacy vulnerabilities in machine learning that grew from a college project to 400+ GitHub stars, 10k+ downloads, and citations in 10+ papers and books. It aims to be a unified API covering a wide range of defenses (differential privacy implementations like DPSGD, AdaDPS, DPlis; Paillier-based homomorphic encryption; k‑anonymity; federated learning algorithms such as FedAvg, FedProx, SplitNN) and attacks (evasion, poisoning, model inversion, membership inference, backdoors). That breadth — rather than single-purpose libraries — makes AIJack useful for running combined experiments, benchmarking defenses against multiple threat models, and teaching ML security concepts reproducibly. The author credits growth to strategic promotion (Papers with Code as the top traffic source, selective Reddit posts, limited Hacker News impact) and rigorous engineering practices: Sphinx docs, Codacy checks, black/isort formatting, pytest/googletest, and GitHub Actions for CI/CD. Those investments lowered the contributor barrier and amplified adoption. Beyond the code, AIJack illustrates how well-packaged OSS can accelerate research, improve interview prospects, and seed academic impact — a practical blueprint for students and researchers who want to build impactful tools in ML security and privacy.