Yaqing Zhou

CMU M.S. (’26), focused on backend systems, data infrastructure, and security.

Email: yaqingz@andrew.cmu.edu

Selected Work

Built an end-to-end club-level fraud/abuse detection service for a ~4M-user gaming loyalty platform (~4M users).
Designed data pipelines, graph-based features (shared founders, member reuse, IP/device overlap, name similarity), and an XGBoost model with semi-supervised label propagation.
Achieved ~3× coverage (190 → 575 clubs) while maintaining <1% FPR and ~0.89/0.89 precision/recall on a held-out test set.
Integrated explainability outputs to speed up manual triage and reduce incentive abuse costs.

Designed a DNS outsourcing protocol allowing external name servers to manage zone files without revealing internal network details.
Implemented reference authoritative and recursive DNS servers in Python (dnspython) and evaluated latency, storage overhead, and leakage.
Measured ~44% lower median CNAME latency vs an encrypted baseline; DNSSEC-related latency overhead <8% vs BIND-9.

Implemented core database components in C++: disk-based storage, LRU-K buffer pool, and concurrent B+Tree index.
Built a query execution engine (scans, joins, aggregation, external sort, basic optimization rules).
Added snapshot-isolation MVCC with per-transaction undo logs, conflict detection, and GC; achieved Top-10/Top-20 performance in course benchmarks.

Completed CTF-style labs in crypto and web security (e.g., SHA-256 length extension, padding-oracle, SQLi/XSS/CSRF) and practiced secure coding by documenting mitigations such as parameterized queries, output encoding/CSP, CSRF tokens, and secure cookie flags.
Took graduate-level security courses at CMU (Security in Networked Systems, Intro to Information Security, Secure Coding), focusing on threat modeling, network/protocol attacks, and secure-by-design systems.
Solved a wide range of CTF challenges (stack/heap overflows, format-string bugs, integer and pointer vulnerabilities, canary bypasses, and blockchain exploits such as insecure randomness and reentrancy).