Latest Interviews

Torch Tutorial (Alex Wiltschko, Twitter)

Alex Wiltschko, Hugo Larochelle, Andrej Karpathy, Richard Socher, Sherry Moore, Ruslan Salakhutdinov, Andrew Ng, John Schulman, Pascal Lamblin, Adam Coates, Quoc Le, Yoshua Bengio, Shubho Sengupta

This presentation details the practical implementation and theoretical foundations of the Torch deep learning framework using the Lua language, developed in collaboration with experts from Facebook, Google, and Twitter. The speaker explains how Torch leverages LuaJIT for high-performance embedded deployment while utilizing its dynamic Autograd system to support flexible control flow and custom gradients without the overhead of static computation graphs. Case studies from Twitter demonstrate the framework's transition from a research tool for cutting-edge models like GANs to a production environment for serving media, highlighting its efficiency in both training via reverse-mode differentiation and inference through lightweight C++ integration.

Deep Learning for Natural Language Processing (Richard Socher, Salesforce)

Richard Socher, Hugo Larochelle, Andrej Karpathy, Sherry Moore, Ruslan Salakhutdinov, Andrew Ng, John Schulman, Pascal Lamblin, Adam Coates, Alex Wiltschko, Quoc Le, Yoshua Bengio, Shubho Sengupta

This presentation outlines the evolution of Natural Language Processing from hierarchical linguistic analysis to deep learning architectures like Word2Vec, GRUs, and Dynamic Memory Networks that handle sequence modeling and visual question answering. Key researchers discussed how continuous vector representations and gated units address ambiguity and long-range dependencies, achieving state-of-the-art performance on benchmarks such as Facebook's bAbI dataset and reducing language modeling perplexity to 70. Despite these advances, the discussion highlights persistent challenges regarding unified joint models, data scarcity in specialized domains, and the need for improved robustness against adversarial inputs and false premises.

Deep Reinforcement Learning (John Schulman, OpenAI)

John Schulman, Hugo Larochelle, Andrej Karpathy, Richard Socher, Sherry Moore, Ruslan Salakhutdinov, Andrew Ng, Pascal Lamblin, Adam Coates, Alex Wiltschko, Quoc Le, Yoshua Bengio, Shubho Sengupta

This technical presentation delineates Deep Reinforcement Learning as a sequential decision-making framework that employs neural networks to maximize cumulative rewards through policy gradients and Q-function learning. Key figures in the field, such as those at DeepMind, have leveraged these methods to master complex environments including Atari games, Go, and robotic locomotion by addressing challenges like reward sparsity and non-stationary state dynamics. The discussion further contrasts algorithmic trade-offs between sample efficiency and robustness while outlining future directions like hierarchical structures and model-based approaches to enhance real-world deployment.

Deep Learning for Computer Vision (Andrej Karpathy, OpenAI)

Andrej Karpathy, Hugo Larochelle, Richard Socher, Sherry Moore, Ruslan Salakhutdinov, Andrew Ng, John Schulman, Pascal Lamblin, Adam Coates, Alex Wiltschko, Quoc Le, Yoshua Bengio, Shubho Sengupta, lexfridman

This presentation traces the evolution of convolutional neural networks from 1960s neuroscience foundations to the 2012 AlexNet breakthrough, highlighting how deep learning displaced traditional feature extraction by achieving near-human accuracy on the ImageNet dataset. Key architectural innovations, such as residual skip connections in ResNets and efficient Inception modules, enabled the training of deeper, wider models that serve as generic feature extractors for diverse tasks ranging from object detection to medical imaging. The discussion concludes with practical deployment strategies emphasizing the use of pre-trained models and GPU-accelerated infrastructure to overcome computational bottlenecks in both cloud and edge environments.

Theano Tutorial (Pascal Lamblin, MILA)

Pascal Lamblin, Hugo Larochelle, Andrej Karpathy, Richard Socher, Sherry Moore, Ruslan Salakhutdinov, Andrew Ng, John Schulman, Adam Coates, Alex Wiltschko, Quoc Le, Yoshua Bengio, Shubho Sengupta

This presentation details Theano, an eight-year-old symbolic expression compiler that enables high-performance deep learning by automatically differentiating mathematical graphs and compiling them into optimized C++ or CUDA code. The session demonstrates the framework's core capabilities, including graph manipulation for neural network backpropagation, GPU acceleration via shared variables, and sequence modeling through the `scan` operator, while showcasing practical implementations of logistic regression, convolutional networks, and LSTMs on datasets like MNIST. Addressing deployment challenges inherent in its tight Python integration, the discussion concludes by highlighting Docker containers as the standard solution for distributing models and outlines a roadmap for enhanced 3D convolution and cuDNN support.

Foundations of Deep Learning (Hugo Larochelle, Twitter)

Hugo Larochelle, Andrej Karpathy, Richard Socher, Sherry Moore, Ruslan Salakhutdinov, Andrew Ng, John Schulman, Pascal Lamblin, Adam Coates, Alex Wiltschko, Quoc Le, Yoshua Bengio, Shubho Sengupta

This lecture by Hugo Lavachel delivers a comprehensive technical analysis of feedforward neural networks, detailing the mathematical foundations of backpropagation, universal approximation, and modern optimization algorithms like Adam and stochastic gradient descent. The presentation critically evaluates architectural choices, contrasting activation functions such as ReLU against sigmoid and tanh while explaining how dropout and batch normalization mitigate overfitting and stabilize training in deep architectures. By bridging theoretical constraints with practical debugging strategies, the course equips practitioners with the necessary tools to implement, tune, and validate deep learning models effectively.

Why OCaml

Approximately a decade ago, Jane Street adopted OCaml as its primary functional programming language to secure C-like performance and rigorous type safety while maintaining development conciseness. This strategic choice serves as a filtering mechanism for hiring, attracting highly motivated engineers through a specialized one-month training program that results in a consolidated, high-performing workforce. Although the firm faces ecosystem limitations compared to mainstream languages, the implementation of a single language across its multi-billion-dollar systems enables significant code reuse and shifts the priority from empirical testing to the universal guarantees of static type checking.

Effective ML 2011 Harvard CS51 Part 1

Ron Minsky

Ron Minsky, a technical leader at Jane Street Capital, presented a comprehensive case for prioritizing code correctness and maintainability through strict OCaml design principles that leverage the type system to enforce data invariants and ensure pattern match exhaustiveness. The lecture outlined specific architectural strategies, including mandatory interfaces, uniform naming conventions, and a "reader-over-writer" philosophy, which collectively minimize cognitive load and prevent silent failures in high-stakes financial systems. Additionally, the speaker addressed immediate course logistics by extending the Moogle project description deadline and recommending simple list-based implementations to avoid unnecessary debugging complexity.

Caml Trading

Carl Curry, Ron Minsky

Jane Street Capital, a proprietary trading firm operating in four global cities, transitioned its primary development stack to the functional programming language OCaml to satisfy critical demands for absolute correctness, high performance, and code maintainability. This adoption enables the firm's 35 core developers to leverage OCaml's powerful type system and algebraic data types for static exhaustiveness checks, ensuring the rigorous standards required to process billions of dollars in daily equity trades without error. While the language presents challenges in ecosystem maturity, the firm has actively mitigated these gaps through its Jane Street Summer Project and achieved superior hiring outcomes by attracting engineers skilled in functional paradigms.

Seven Implementations of Incremental

Incremental is an internal OCaml library that efficiently refreshes large computations by modeling logic as a dependency graph, allowing localized updates when only small data subsets change. Through eight iterative development phases, the team resolved critical academic limitations regarding garbage collection and dynamic graph restructuring, culminating in a production-ready V8 implementation that achieved a threefold speedup using generalized algebraic data types. This technology successfully transformed a slow trading system frontend into the suite's fastest application and is now being explored for reactive JavaScript compilation and efficient functional data structure diffing.

MI Forum: Modern-Day Magellan: JPL Director Charles Elachi

Charles Elachi

Dr. Charles Elachi, Director of NASA's Jet Propulsion Laboratory, detailed the successful "Seven Minutes of Terror" landing of the Curiosity rover, which has confirmed the presence of essential chemical elements and ancient flowing water on Mars. The presentation highlighted the mission's scientific outcomes, including the discovery of non-oxidized surface material and the approval of future seismology and sample-return missions, while noting that the rover's nuclear power system is designed to operate for a decade. Elachi also projected that human exploration of Mars could commence within 20 to 25 years, citing technological advancements in fuel production and the ongoing development of the James Webb Space Telescope.

Aubrey de Grey: "The cost effectiveness of anti-aging research"

Aubrey de Grey, Ben

Aubrey de Grey proposes a "maintenance" strategy within the SENS framework to systematically repair seven categories of cellular damage, challenging the inevitability of aging by targeting its biological root causes rather than treating individual diseases. This approach aims to achieve "longevity escape velocity," a threshold where therapeutic advances outpace degeneration to eventually reduce mortality from aging to negligible levels comparable to accidents. By shifting billions in global healthcare spending toward preventative rejuvenation, the strategy promises to extend healthy lifespans significantly while mitigating economic costs and addressing ethical concerns regarding overpopulation and identity.

California's Stem Cell Initiative: Mapping the New Frontier of Medicine

Dr. Alan Trounson

Scheduled for April 27–29 at the Beverly Hilton Hotel, the upcoming global conference will feature strategic updates from CIRM leadership, including discussions on $632.5 million in awarded grants and a new billion-dollar commitment to Disease Teams. The meeting also addresses critical logistical changes, such as a 10% attendance reduction to improve quality, alongside presentations on overcoming funding gaps caused by California's budget impasse. With key figures like Chris Ailman participating in related February events, the agenda emphasizes accelerating translational breakthroughs in stem cell therapies for conditions like spinal cord injury and diabetes while navigating complex regulatory landscapes.