Lecture, Conference Presentation
Caml Trading
- Jane Street anticipates maintaining daily trading volumes of 200–400 million shares and $4–8 billion in nominal value, necessitating software capable of millisecond-level reactions to 1TB of daily raw data.
- The firm expects the financial landscape to evolve rapidly due to new regulations and securities, requiring codebases that can be updated efficiently to adapt to the changing environment.
- Funding for the Jane Street Summer Project will continue and expand to include applications from other functional programming communities, with a scope growth targeting user interfaces and library support beyond core languages.
- Plans include bridging the gap between academic research and practical tooling by funding students to work with faculty, despite OCaml not being a natural incentive for academia.
- The firm intends to recruit high-quality programmers capable of learning ML through training, prioritizing the ability to teach talent over hiring established experts.
- Anticipated technical evolution involves a shift toward Windows-agnostic high-performance trading systems to ensure precise timekeeping, alongside a move toward writing code that operates across multiple processes.
- The industry faces a predicted need for well-architected infrastructure to manage concurrency across multiple processors and address challenges in "programming in the large," such as package distribution and namespaces.
- Concerns exist that a lack of robust concurrency infrastructure and library support could become a bottleneck as the industry increases parallelism.
- OCaml is expected to remain preferred over F# due to F# current performance limitations with symbolic code, high allocation rates, and the persistence of null reference risks inherent in .NET libraries.
- High-performance code remains essential to avoid excessive hardware costs and complexity management, given the premium placed on human time in trading operations.