Programming Language Engineer

• Support for high-performance computing, including work on tooling, verification techniques, and languages for hardware design, as well as support for programming GPUs and other ML accelerators.
• Domain-specific languages. Designing, implementing, and evolving custom languages that power some of our most important internal platforms.
• Tooling and developer experience. Profilers, debuggers, language servers, IDE integrations, documentation tools, and the surrounding infrastructure that makes a language genuinely pleasant to use at scale.

Some of this work involves working on open-source language implementations, like those of OCaml and Python, and for those areas, a meaningful portion of our work is open-source as well. Some of it involves custom languages and DSLs that we’ve built entirely in-house, some open-source, some purely internal.

• Deep experience with-and love for-programming languages. There’s no specific checklist; we are interested in people with experience in language design, type systems, compilers, runtimes, or DSL implementation
• Top-notch programming skills
• Strong interpersonal skills. Software engineering at Jane Street is highly collaborative, and we are looking for people who can work effectively in small, close-knit teams

No previous experience with OCaml or functional programming is required.

We’re looking for Software Engineers with a background in programming languages to join one of our PL-focused teams at Jane Street.

Programming languages work at Jane Street spans a wide variety of different areas:

• Language design and type systems. Designing language features and stronger type-level guarantees that let our developers express more of what they mean, and catch more mistakes at compile time.
• Compilers and optimization. Investing in optimization, code generation, and backend work that translates directly into measurable speedups across the firm’s most performance-sensitive systems.
• Incremental and parallel computation. Building runtimes that can evaluate large, irregular, dynamic computation graphs incrementally and/or parallel across many cores and machines, with first-class attention to scheduling, memory, and resource accounting.