Resume

JOHN QUINN

Compilers and Database Engine Development | Rust, Java, Kotlin, C, C++
Email: [email protected] | GitHub: johnedquinn
LinkedIn: johnedquinn

Profile

Compiler and database engine engineer specializing in query language implementation and performance. At Amazon, designed PartiQLVM, a register-based bytecode VM in Rust achieving ~20x latency improvement over the prior implementation, and led the v1.0.0 release of PartiQL powering multiple production AWS services. Equally comfortable in systems languages (Rust, C, C++) and the JVM (Kotlin, Java), and drawn to work where language design, compiler theory, and performance engineering intersect.

Experience

Amazon SDE 1 (2021 — 2023); SDE 2 (2023 — present)

Amazon: PartiQL (Rust)

• Designed and built PartiQLVM, a register-based bytecode virtual machine in Rust, achieving ~20x latency improvement over the previous Rust implementation through zero-copy data analysis, a cacheable/thread-safe compiled bytecode program, and projection pushdown for non-relational data. (v0.15.0-alpha.1)
• Authored partiql-jni, a zero-dependency Java library wrapping PartiQLVM via JNI bindings; designed a BufferedCatalog with a custom binary encoding to minimize serialization overhead across the Java/Rust boundary, delivering a 2-3x latency improvement over partiql-lang-kotlin.

Amazon: PartiQL (JVM)

• Led the v1.0.0 release of partiql-lang-kotlin, an open-source SQL-compatible query language powering multiple Amazon and AWS services. (v1 Release)
• Architected the plan-based evaluator engine from scratch (PR #1283); implemented core relational operators including JOINs (PR #1294), aggregations, correlated subqueries, set operations, and dynamic function dispatch with type-directed optimization.
• Designed and implemented the PType (PR #1488) and Datum (PR #1451) APIs forming the runtime type system.
• Designed and implemented the error reporting framework (PR #1615) and PERMISSIVE/STRICT evaluation modes (PR #1353) adopted across the production compiler pipeline.
• Implemented Common Table Expressions (CTEs) end-to-end across parsing, planning, and evaluation; added SQL-standard interval data types supporting arithmetic over YEAR, MONTH, DAY, HOUR, and finer-grained units. (v1.1.0, v1.2.0)
• Implemented window functions (RANK, DENSE_RANK, ROW_NUMBER, LAG, LEAD) and the WINDOW clause operator, enabling inter-row analytical queries and eliminating the need for CPU-intensive self-joins. (PR #1746, v1.3.0 Release)
• Authored the PartiQL parser using ANTLR4: wrote the G4 lexer/grammar (~1.1K lines), implemented a parse tree visitor converting ANTLR nodes to the PartiQL AST, and resolved 6 pre-existing parser correctness bugs. (PR #711)

Projects

Kanonic

• Designed and implemented a fast LR(1) parser generator for the JVM, written in Kotlin. Implements canonical LR(1) item set construction, conflict resolution, and code generation targeting the JVM — motivated by the lack of a performant JVM-native parser generator suitable for production use.

Education

B. Engineering, Computer Science — University of Notre Dame ‘21