Latest Interviews
Showing 16–30 of 48 transcripts.
Clear all filters- Jane Street41 min
Charlie Marsh: uv: An Extremely Fast Python Package Manager
Charlie Marsh of Astral introduces UV, a Rust-based Python package manager that functions as a unified drop-in replacement for tools like pip and Poetry by combining installation, dependency resolution, and environment management into a single high-performance workflow. Since its mid-February release, the tool has achieved 16 million monthly downloads by addressing Python's complex dependency graph challenges through advanced SAT solving, Algebraic Decision Diagrams, and a zero-copy architecture that treats virtual environments as ephemeral. UV now commands over 10% of all PyPI requests, leveraging custom version parsing and cache reflinking to enable near-instant package installation and declarative dependency locking.
- Jane Street1h 6m
The Uncertain Art of Accelerating ML Models with Sylvain Gugger
Sylvain Gugger, Ron Minsky, Jeremy Howard, Mark Mandelmann, Mark Mirchandani, Francesc Campoy, Gabriel Sanchez
Former fast.ai co-author Jeremy Howard discusses his transition from mathematics education to optimizing machine learning infrastructure at Jane Street, highlighting breakthroughs in learning rate schedules and image resizing that previously secured top benchmark placements. He details the development of the Hugging Face Accelerate library, a lightweight tool designed to abstract complex hardware parallelism and eliminate boilerplate code for training across diverse GPUs and TPUs. The discussion further explores the architectural constraints of financial data, the dominance of PyTorch's iterative execution model, and Jane Street's rigorous approach to reproducibility and custom model development for high-frequency trading.
- Jane Street1h 12m
State Machine Replication, and Why You Should Care with Doug Patti
Doug Patti's presentation details Jane Street's architectural evolution from the rigid, single-sequencer Concord system to the more flexible ARIA infrastructure, which supports 24/7 operations and hierarchical topic namespaces. ARIA addresses previous limitations in versioning and overhead while preserving global deterministic ordering for high-frequency trading, enabling internal teams to manage complex workflows with reduced operational burden. Despite these advancements, the architecture still faces challenges regarding rate limiting and snapshot capabilities, prompting ongoing optimizations to support broader firm-wide adoption.
- Jane Street1h 4m
Building a Functional Email Server with Dominick LoBraico
Faced with the security risks and operational rigidity of a legacy mail server, Jane Street engineers replaced the system with MailCore, a homegrown server implemented in OCaml where configuration is expressed as composable code. To ensure safety during the one-year migration, the team deployed a parallel shadow instance that continuously diffed outputs against the legacy system, effectively reverse-engineering implicit knowledge while eliminating memory-safety bugs common in C-based alternatives. This architectural shift transformed email management into a software culture practice, allowing generalist engineers to safely modify routing and filtering logic using standard OCaml tools rather than relying on specialized knowledge of opaque custom domains.
- Jane Street1h 5m
Raph Levien: A Taste of GPU Compute
The presentation details how modern GPU architectures prioritize throughput over single-thread latency to overcome the plateau of CPU scaling, leveraging parallelism to achieve tenfold improvements in performance per dollar and watt. It outlines a functional programming model for GPU compute that employs hybrid tiling and bitmasking strategies to minimize branch divergence and maximize read efficiency in rendering workloads. Finally, the event analyzes the industry's shift from proprietary CUDA ecosystems toward standardized open APIs like Vulkan and WebGPU, which facilitate portable, high-performance computing through intermediate languages like SPIR-V and formal memory models.
- Jane Street58 min
Matt Might: The Algorithm for Precision Medicine
Matt Mite, director of the Hugh Call Precision Medicine Institute at UAB, leverages computational optimization and a knowledge graph of 30 million abstracts to derive actionable precision medicine strategies for patients with undiagnosed genetic disorders. His approach, exemplified by the successful treatment of his son Bertrand's NGLY1 deficiency and the identification of Prevacid for ion channel epilepsies, prioritizes logical mechanism proofs to overcome physician skepticism and legal liability. Through an undergraduate-led workflow and the "All of Us" initiative, Mite's institute continues to transform complex genomic data into targeted interventions, from drug repurposing to novel metabolic pathways, asserting that scientific experimentation always yields actionable solutions.
- Jane Street1h 24m
Steve Lee: State of the Shell: PowerShell7
The PowerShell team announced a strategic shift where Windows PowerShell 5.1 is now maintenance-only, while the primary focus moves to PowerShell 7 as a cross-platform Long-Term Servicing release aligned with the .NET yearly cycle. Significant technical improvements include the introduction of null propagation, parallel processing capabilities, and a new module management system designed to replace the deprecated v2 architecture. While legacy Windows-specific modules now access via implicit remoting, the product strategy excludes future in-box inclusion in Windows due to mismatched support lifecycles.
- Jane Street50 min
Michael Sikorski: Dude, Where Are My Files? Reverse Engineering Ransomware
FireEye malware analyst Michael Sikorski details the escalating threat landscape of financially motivated ransomware, highlighting catastrophic real-world impacts on organizations like Maersk and the shift of state actors toward direct extortion. He outlines essential analysis methodologies, ranging from static binary examination to deep reverse engineering, which revealed critical kill switches for devastating strains such as WannaCry and the destructive EternalPetya attack. Ultimately, the presentation emphasizes how these technical insights drive the industry's response to complex encryption schemes and enable data recovery assessments or containment measures against rapidly evolving malicious code.
- Jane Street47 min
Unboxed Types for OCaml
The presentation "Making a Camel Less Like Lisp and More Like C++" addresses OCaml's performance limitations by introducing explicit "layout kinds" that allow the compiler to statically determine memory representations for unboxed types like 64-bit integers and records. This system distinguishes between tagged pointers managed by the garbage collector and raw bit representations passed directly in registers, enabling efficient libraries to remain opaque while avoiding the overhead of standard boxing. By replacing fragile dynamic heuristics with explicit type annotations, the proposal facilitates predictable high-performance computing without sacrificing the language's safety guarantees or requiring complex template expansion.
- Jane Street1h 6m
Derek Dreyer: RustBelt: Logical Foundations for the Future of Safe Systems Programming
Derek Dreyer, Ralf Jung, Jacques-Henri Jourdan, Robbert Krebbers, Hoang-Hai Dang, Jan-Oliver Kaiser
The MPI for Software Systems' "Rust Belt" project applies decades of academic research to formally verify the safety guarantees of the Rust programming language against the risks posed by "unsafe" code. By leveraging the Iris framework and mechanizing proofs in the Coq assistant, the team establishes a semantic safety model that independently verifies complex libraries like `Arc` and `Mutex`, successfully identifying critical bugs in memory ordering and reference counting. This methodology ensures that future language evolution and standard library modifications can proceed with mathematical certainty that undefined behavior remains impossible.
- Jane Street1h 13m
Playing Atari Games in OCaml
This presentation explores the application of machine learning to Atari games by leveraging OCaml's functional type safety alongside PyTorch and TensorFlow for GPU-accelerated deep learning. It details how supervised convolutional neural networks achieve over 99.5% image classification accuracy and how reinforcement learning agents utilizing Deep Q-Learning can master games like Breakout and Pong, albeit with significantly slower training times than humans. The talk concludes by highlighting the technical interoperability achieved through custom OCaml bindings and outlines future research directions in transfer learning and generalization across varying game environments.
- Jane Street1h 33m
Don Jones: 13 Years in a Shell: Lessons, Practices, and Achievements in PowerShell
Don Jones, Jeffrey Snover, Arnold, Doug Fake
Hosted by Arnold and Doug Fake, the New York PowerShell community meetup featured veteran practitioner Don Jones discussing his evolution from Navy maintenance to modern DevOps strategy. Jones emphasized critical technical anti-patterns such as avoiding Format-* cmdlets and global variables while advocating for Test-Driven Development, Git integration, and the Single Responsibility Principle to ensure script maintainability. The session concluded with guidance on organizational alignment, noting that true DevOps requires cross-functional product teams and continuous skill investment to remain competitive in a fast-moving market.
- Jane Street58 min
Safe at Any Speed: Building a Performant, Safe, Maintainable Packet Processor
Jane Street engineers optimized their OCaml-based market data distribution system to handle NASDAQ's peak load of 4 million messages per second while maintaining zero-allocation on critical paths to avoid garbage collection delays. By leveraging PPX preprocessors, immediate integer options, and a domain-specific language for protocol generation, the team reduced per-message processing latency from five microseconds to under 750 nanoseconds. This approach demonstrates that strict single-core, low-latency performance targets can be achieved with high-level functional languages through aggressive inlining and careful memory management rather than resorting to lower-level systems code.
- Jane Street1h 9m
A Language-Oriented System Design
The Ontology system presents a dual-language architecture that manages organizational permissions through an English-facing interface backed by a dependently typed, deductive database engine. It distinguishes between hard-coded primitives, extensional tables, and intentional logical rules to enforce strict type safety, ensuring that data retrieval returns verifiable proof objects and automatically maintains referential integrity through cascading deletion and trash mechanisms. While the internal Prolog-like inference engine mitigates undecidability risks via execution budgets, the design allows for extensible permission modeling by uploading new type declarations without requiring code-level system upgrades.
- Jane Street1h 18m
Lindsey Kuper: Abstractions for Expressive, Efficient Parallel and Distributed Computing
Lindsey Kuper, Ron, Laura, Kunle Olokutun, Asante Toney, Neil Conway, Peter Alvaro, Carl, Nikki Vazou
The presentation outlines a unified framework for deterministic parallel computing that utilizes Lattice Variables (LVARs) and Conflict-Free Replicated Data Types to ensure correct distributed executions while enabling high-performance acceleration in Julia and Python through non-invasive domain-specific languages. Complementing these runtime systems, the speaker details a formal verification approach for safety-critical neural networks that employs lazy ReLU splitting within SMT solvers to manage the exponential complexity of non-linear activations. Future research aims to democratize solver development by integrating lattice theories directly into SMT architectures and verifying LVAR constraints through advanced type systems.