Machine Learning Research Scientist - Interpretability

Join Duolingo, the world's leading language-learning platform, as the Principal Learning Scientist and Head of Efficacy Research. In this pivotal role, you will spearhead innovative research initiatives to assess and enhance the efficacy of our educational products. Your expertise will guide the development of evidence-based strategies that ensure our learners achieve their language goals effectively.As the Head of Efficacy Research, you will collaborate with cross-functional teams, including product development and data analytics, to design robust research methodologies. You will analyze user data and feedback, translating insights into actionable recommendations that drive product improvements. This is an exciting opportunity to influence the future of language education on a global scale.

About BasisBasis is a pioneering nonprofit organization dedicated to applied AI research. Our mission is twofold: to deepen our understanding of intelligence and to empower society in solving complex, challenging problems.Our first objective is to comprehend and construct intelligence. This involves establishing foundational mathematical principles for reasoning, learning, decision-making, comprehension, and explanation, while also developing software that embodies these principles.Secondly, we aim to enhance society’s capability to tackle intractable issues. This includes broadening the scope and complexity of problems we can address today and accelerating our future problem-solving abilities.To accomplish these objectives, we are creating a technological infrastructure inspired by human reasoning and fostering a collaborative organization that prioritizes human values.About the RoleAs a Research Scientist, you will spearhead Basis’ initiatives to unravel the conceptual, mathematical, and computational foundations of intelligence. We seek individuals with exceptional technical skills who are eager to explore foundational concepts. Our research scientists and engineers are committed to conducting rigorous, high-quality scientific investigations, while also embracing the opportunity to experiment, learn from errors, and explore innovative ideas.Collaboration is integral to our work, both internally and with external partners. We are searching for individuals who thrive in team settings, tackling challenges that transcend individual capabilities.Key Expectations:Proven ability to conduct high-quality scientific research, demonstrated through publications, technical reports, or software projects.Preferred Qualifications:PhD or equivalent experience in relevant fields such as statistics, programming languages, machine learning, computational neuroscience, cognitive science, physics, or mathematics.Enthusiasm for addressing real-world challenges and contributing to positive societal change.ResponsibilitiesDevelop and investigate computational theories of intelligence, encompassing reasoning, learning, and decision-making.Collaborate closely with domain experts, both within Basis and externally, to contribute to solving scientific problems.

About BasisBasis is a nonprofit organization dedicated to applied AI research, striving to achieve two interconnected objectives.Firstly, we aim to understand and develop intelligence. This involves establishing the mathematical foundations of reasoning, learning, decision-making, comprehension, and explanation; along with creating software that embodies these principles.Secondly, we seek to enhance society's capacity to address complex challenges. This means broadening the range, scale, and intricacies of the problems we can tackle today, while also accelerating our ability to solve future challenges.To fulfill these missions, we are constructing a novel technological framework inspired by human reasoning, and fostering a collaborative organization that prioritizes human values.About the RoleAs a Research Scientist, you will spearhead Basis's initiatives to deepen our comprehension of the theoretical, mathematical, and computational aspects of intelligence.We seek individuals with exceptional technical skills who are passionate about exploring concepts at their core. Our research scientists and engineers are committed to conducting rigorous, high-quality, and robust scientific inquiry, yet they embrace experimentation, learning from mistakes, and exploring innovative ideas to achieve their goals.Basis thrives on collaboration, both internally and with external partners; thus, we value team players who relish tackling challenges that surpass individual capabilities.Research FocusOur research under the MARA project is dedicated to forging new principles and technologies for modeling, abstraction, and reasoning in AI systems. The primary aim of MARA is to reveal principled methodologies for how intelligence constructs, refines, and applies world models through interactive experimentation.For this position, we specifically seek experts in Reinforcement Learning & Planning who can push the boundaries of model-based RL, exploration strategies, optimal control, and Bayesian optimization. You will develop agents capable of learning efficient policies in intricate, partially observable environments by utilizing structured world models.

About BasisBasis is a nonprofit organization dedicated to applied artificial intelligence research. Our mission is twofold: to understand and build intelligence and to advance society’s problem-solving capabilities. We strive to unravel the mathematical principles behind reasoning, learning, decision-making, understanding, and explanation, while also developing software that embodies these principles.Our commitment extends to enhancing our ability to tackle complex issues that are beyond today's capabilities and accelerating our potential to address future challenges. We are creating a technological framework inspired by human reasoning, alongside fostering a collaborative organization that prioritizes human values.About the RoleAs a Research Scientist, you will spearhead Basis’ initiatives to deepen our understanding of the conceptual, mathematical, and computational principles of intelligence. We seek individuals who excel technically and are passionate about exploring foundational concepts. Our research team values rigorous, high-quality scientific endeavors while encouraging experimentation and exploration of innovative ideas.Basis thrives on collaboration, both internally and with external partners. We are looking for team players who enjoy tackling significant problems that require collective effort.Research FocusDespite the growing acknowledgment that acquiring and understanding world models is crucial for intelligence, current AI systems face challenges in mirroring this human capability. Key uncertainties remain regarding the essence of a world model, methods for reliably detecting its presence in agents, and approaches to develop agents capable of learning these models effectively.Our research, particularly within the MARA project, seeks to establish new foundations and technologies for modeling, abstraction, and reasoning in AI systems. MARA's goal is to identify principled methods for how intelligence constructs, refines, and employs world models through interactive experimentation. Achieving this will require advancements in knowledge representation, abstraction, reasoning, active learning, and reinforcement learning, necessitating a first-principles reevaluation of world modeling.

Become a pivotal part of our innovative team at Output Biosciences, where you will contribute to the groundbreaking development of the world's inaugural biological reasoning model. Collaborate with us in creating generative foundational models that decode biological systems at all scales—from molecules to entire organisms—allowing us to predict, understand, and manipulate living systems in unprecedented ways.Currently operating under stealth mode, Output Biosciences is spearheaded by a team of accomplished founders and seasoned biotech veterans with a proven track record of successful exits in both AI and biology. We are proudly backed by leading venture capital firms including Y Combinator.As a Machine Learning Research Scientist specializing in Generative AI, you will work closely with our founders and collaborative team members to forge advanced generative models capable of synthesizing and designing biological systems across various scales. Building on our biological reasoning model, your focus will be on creating generative capabilities that translate insights into innovative applications.Design and develop state-of-the-art generative models tailored for biological data, enabling AI systems to generate novel molecular structures, protein designs, cellular components, and biological pathways.Implement cutting-edge generative algorithms including diffusion models, language diffusion models, diffusion transformers, and flow matching approaches, customized for biological constraints and properties.Train and scale pioneering AI models utilizing large, diverse multimodal biological datasets across distributed GPU clusters.Solve complex AI research challenges at scale, addressing crucial aspects like model architecture, scaling, pre-training and fine-tuning strategies, and rigorous evaluation methods.Explore and implement various advanced generative AI architectures, such as transformers, diffusion models, and autoencoders, to push the boundaries of biological AI innovation.

About BasisBasis is a pioneering nonprofit organization dedicated to applied AI research, driven by a dual mission.Firstly, we aim to deepen our understanding and development of intelligence. This involves establishing the mathematical foundations of reasoning, learning, decision-making, understanding, and explanation, along with creating software that embodies these principles.Secondly, we strive to enhance society’s capacity to tackle complex challenges. This means broadening the scale and complexity of the problems we can address today, while also accelerating our future problem-solving capabilities.To realize these aims, we are constructing an innovative technological framework inspired by human reasoning and fostering a collaborative organization that prioritizes human values.About the RoleAs a Research Scientist at Basis, you will play a crucial role in advancing our understanding of the theoretical, mathematical, and computational principles underlying intelligence.Our Research Scientists possess key characteristics:Outstanding technical expertise—strong mathematical and computational foundations.A creative builder’s mindset—capable of designing, constructing, and refining complex systems based on foundational principles.Commitment to scientific rigor—engaging in high-quality, robust scientific inquiry without hesitance to experiment, learn from mistakes, and explore unconventional ideas.At Basis, collaboration is key, both within our teams and with external partners. We seek individuals who thrive in collaborative environments and are eager to tackle challenges that transcend individual capabilities.Programming Languages Research ScientistsThis role is tailored for experts in the design, implementation, and analysis of programming languages. You will assist in the design and implementation of the foundational computational reasoning systems under development at Basis.Key focus areas of programming languages research include compiler design, partial evaluation, program analysis, abstract interpretation, and program transformation. This work is conducted within the context of constructing reasoning systems, leading to engagement with topics such as probabilistic programming, automatic differentiation, and SAT/SMT solvers.Expectations:Possess a PhD (or equivalent experience) in a relevant field.

About BasisBasis is a pioneering nonprofit organization dedicated to applied AI research, driven by dual objectives that enhance each other.Our primary focus is to understand and construct intelligence. This encompasses establishing the fundamental mathematical principles of reasoning, learning, decision-making, comprehension, and explanation, alongside developing software that embodies these principles.Secondly, we aim to empower society to tackle complex challenges. This involves broadening the scale, complexity, and scope of problems we can currently address and, crucially, accelerating our future problem-solving capabilities.To realize these ambitions, we are establishing an innovative technological framework inspired by human reasoning and fostering a collaborative environment that prioritizes human values.About the RoleAs a Research Engineer, you will play a vital role in advancing Basis’ mission by converting research concepts into accurate, robust, and scalable high-quality code.We are looking for technically proficient individuals who are enthusiastic about delving deep into foundational concepts. Our research engineers are committed to conducting rigorous, high-quality science, unafraid to experiment, learn from mistakes, and explore innovative ideas.Basis thrives on collaboration, both within our team and with external partners. We seek individuals who enjoy tackling challenges greater than they can manage alone.Machine Learning Research Engineer Focus AreasThis position is aimed at experts in machine learning engineering. Key areas of focus include:Probabilistic programming and statistical inferenceDeep learningCausal inferenceProgram synthesis and analysisML Ops and systems engineeringThese areas will be explored in the context of developing reasoning systems. Research engineers will also engage with topics including programming language design and implementation, automatic differentiation, and SAT/SMT solvers.

About BasisBasis is a pioneering nonprofit organization dedicated to applied AI research, pursuing two interconnected goals.Our primary objective is to understand and create intelligence. This involves formulating the mathematical foundations of reasoning, learning, decision-making, understanding, and explanation, while developing software that embodies these concepts.The second goal is to enhance society’s capacity to solve complex challenges. We aim to broaden the scope, intricacy, and variety of challenges we can address today, and crucially, to expedite our ability to tackle future problems.To realize these objectives, we are constructing an innovative technological base inspired by human reasoning, and fostering a collaborative organizational culture that prioritizes human values.About the RoleAs a Research Scientist on the MARA (Modeling, Abstraction, and Reasoning Agents) initiative, you will be instrumental in developing computational theories of scientific reasoning in the realms of robotics and embodied intelligence. Your work will push the boundaries of world modeling, reinforcement learning, program synthesis, and robotic control, enabling systems to learn, reason about, and engage with the physical world.We are seeking outstanding researchers with a strong background in Program Synthesis & Neuro-symbolic Methods. The ideal applicant will have a robust publication record in esteemed venues, blending theoretical expertise with practical implementation skills, and a passion for building systems that emulate scientific learning—hypothesizing, experimenting, and modeling the workings of the world.In this role, you will collaborate with an interdisciplinary team to address fundamental inquiries: How can agents learn causal models through interaction? How do we connect high-level reasoning with low-level control? How can we produce interpretable, verifiable control programs instead of opaque policies?At Basis, we value collaboration, both within our team and with external partners; we seek individuals who thrive in teamwork on significant challenges that require collective effort.

Who We AreAt Percepta, our mission is to revolutionize essential sectors through applied AI. We are dedicated to ensuring that key industries, such as healthcare, manufacturing, and energy, leverage cutting-edge technology. By collaborating with industry leaders, we facilitate AI transformation.Expertise in engineering, product development, and research.Mosaic, our proprietary toolkit for swift deployment of agentic workflows.Strategic alliances with companies such as Anthropic, McKinsey, AWS, and others in the General Catalyst portfolio.Our team is rapidly expanding, consisting of Applied AI Engineers, Embedded Product Managers, and Researchers who are passionate about integrating AI advancements into our daily lives. Percepta is a direct partnership with General Catalyst, a global leader in transformation and investment.About the RoleAs a Research Engineer/Scientist (Optimization) at Percepta, you will play a crucial role at the convergence of AI research and impactful real-world applications. You will push the boundaries of decision-making in vital industries by integrating state-of-the-art machine learning with robust optimization research. Collaborate closely with our Embedded Product Managers (EPMs) and engineers to ensure that our advanced decision systems are both innovative and practically beneficial for transforming organizational operations.ResponsibilitiesEstablish and lead ambitious research initiatives that broaden the horizons of data-driven decision-making.Develop innovative optimization and machine learning methodologies for significant challenges such as planning, scheduling, routing, pricing, and inventory management.Create high-fidelity simulators and comprehensive benchmarks that accurately reflect real-world constraints, uncertainties, and multi-objective trade-offs.Translate research into practical applications by partnering with engineers to swiftly prototype solutions and effectively implement research insights.You May Be a Great Fit If YouHold a degree in Computer Science, Operations Research, Industrial Engineering, or Applied Mathematics (MS/PhD preferred) or possess equivalent research or industry experience.Have demonstrated experience in optimization and machine learning.Exhibit a strong ability to collaborate and communicate effectively across multidisciplinary teams.

Jump Trading Group is at the forefront of world-class research, empowering talented individuals in Mathematics, Physics, and Computer Science to explore scientific boundaries and apply innovative research to the global financial markets. Our unique culture thrives on constant innovation, which requires creativity, intellectual honesty, and a relentless competitive spirit. We believe in collaborative success, unlocking individual talents through mutual respect and teamwork. At Jump, our research not only drives superior risk-adjusted returns but also leads to the development of transformative technologies that impact the world, funding startups across various industries, and partnering with premier research organizations and universities to tackle complex challenges.Our AI team serves as a pivotal research and development hub, consisting of quantitative researchers, engineers, and machine learning experts dedicated to advancing large language model (LLM) research and its applications. Our mission is to merge cutting-edge techniques with original research to derive actionable insights from unstructured data.

About Mirage Mirage builds an AI-powered video platform that brings production and editing together through natural language. The team focuses on models with strong contextual understanding, aiming to match the creative decisions of skilled editors. This approach boosts productivity for experienced teams and makes advanced video creation accessible to a wider audience. Learn More Product: Captions by Mirage Research: Seeing Voices (technical white paper) Updates: Mirage on X / Twitter Press: TechCrunch, Forbes AI 50, Fast Company Investors Mirage is backed by leading investors and entrepreneurs, including Index Ventures, Kleiner Perkins, Sequoia Capital, Andreessen Horowitz, and General Catalyst. Role Overview: Research Scientist, Video Generation This position is based at Mirage’s headquarters in Union Square, New York City. In-person attendance is required. Mirage seeks a Research Scientist with a focus on video generation. As an early member of the team, this person will help address foundational challenges in generative media. The role involves developing new modeling techniques, designing training methods, and scaling large video models for real-world use. Work in this area shapes systems used by millions of creators. What You Will Do Advance the quality, controllability, and realism of video generation models Focus on areas such as facial expression synthesis and scene understanding Develop and refine modeling and training strategies for large-scale video systems Collaborate with an interdisciplinary team to solve technical and creative challenges in generative media

Join our innovative AI Research team at Datadog as an AI Research Scientist. Collaborate with talented research engineers and work closely with our Product and Engineering teams to transform cutting-edge research into valuable solutions for our customers.At Datadog AI Research, we are proud of our successful AI-driven solutions, including Bits AI, Watchdog, and Toto. Our focus is on tackling high-stakes, ambitious projects that address real-world challenges in cloud observability and security.We are currently concentrating on three pivotal research areas:Observability Foundation Models – Developing advanced models for forecasting, anomaly detection, and comprehensive telemetry analysis across logs, metrics, and traces. These models will empower our agents to analyze telemetry data effectively.

Join Formation Bio as a Principal Data Scientist, where you will play a pivotal role in shaping our data-driven strategies. We are seeking a highly skilled and innovative data scientist to lead complex projects and drive actionable insights from our data.In this position, you will collaborate with cross-functional teams, leveraging your expertise in statistical analysis, machine learning, and data visualization to inform decision-making and improve our products. Your contributions will directly impact our mission to harness data for groundbreaking advancements in biotechnology.

About the Personalization Team Spotify’s Personalization team works to make discovering new music and podcasts easier and more enjoyable. Well-known features like Blend and Discover Weekly rely on a deep understanding of audio content to create recommendations tailored to each listener. These efforts reach millions worldwide, helping people find audio that fits their tastes. Role Overview The research group within Personalization leads advances in machine learning and artificial intelligence. The team explores new methods, questions established approaches, and shares findings with the broader research community. Senior Research Scientists here help set the direction for personalization technology and have a direct impact on Spotify’s product strategy and user experience.

Hello! I'm Brian, Co-Founder of Egra. We are thrilled to announce that we have successfully raised $5.5 million to develop foundational models for interpreting brain signals, and we are eager to welcome our inaugural research scientist to our innovative team.In this role, you will be empowered to take complete ownership of your projects from the outset. Say goodbye to prolonged onboarding processes, the need for excessive approvals, and bureaucratic hurdles. You will collaborate directly with the founders, tackling complex technical challenges and utilizing the resources available to drive solutions. You will have the unique opportunity to shape the research vision, make key architectural choices, and lay the groundwork for our core technology. If you're someone who thrives in an environment that offers high autonomy and desires to make a significant impact on the direction of a pioneering company, this role is perfect for you.Key ResponsibilitiesEEG (electroencephalography) is one of the most challenging real-world signal modalities in machine learning due to its low signal-to-noise ratio, vast subject variability, and inconsistencies across devices. Many shy away from this domain for these reasons.As our founding research scientist, you will collaborate closely with us to discover effective methodologies. Here’s a glimpse into the projects you will spearhead:Developing self-supervised pretraining objectives that enhance generalization across various subjects, devices, and recording conditions.Critically evaluating existing methods (including recent EEG foundation model studies) to identify their limitations across datasets and conditions, particularly under distribution shifts.Creating robust evaluation protocols that effectively differentiate genuine progress from noise, ensuring we maintain rigorous benchmarks.Documenting internal research findings through comprehensive memos that contribute to the lab's collective knowledge — addressing questions like "why model X fails on dataset Y" and "what Z dataset misleads us about".This position is about cultivating a thorough understanding of a modality that remains largely unexplored with modern machine learning approaches and transforming this insight into impactful advancements.Vision for the FutureWe are striving towards a future where thought serves as a direct interface.Imagine composing a message silently as it types itself or navigating an AR environment without any physical interaction. Software that adapts to your cognitive state in real-time. We are developing a universal interface connecting human cognition with digital actions.Our product development encompasses three pivotal layers:A Neural Encoder: a foundational model designed to translate raw EEG data into resilient, reusable embeddings that function seamlessly across diverse devices and recording environments.

Join Integrated Resources Inc. as a Research Scientist I, where you will contribute to groundbreaking scientific discoveries and innovative solutions. This role offers a fantastic opportunity for recent graduates or early-career professionals looking to expand their expertise in a collaborative and dynamic environment.

About ASAPP ASAPP develops AI technology to improve customer experience. Teams work with a focus on customer needs, clear goals, and measurable results. The company values ownership, purposeful speed, and ongoing improvement based on data and craft. ASAPP’s workforce spans New York City, Mountain View, Latin America, and India. The culture supports both hybrid and remote work, bringing together talent from around the world. Role Overview: Research Scientist, Conversational AI The research team is growing and seeks an experienced Research Scientist with expertise in Conversational AI and Machine Learning. This role focuses on solving scientific problems in Conversational AI and building domain-specific, real-world applications. The position involves working with large datasets from ASAPP’s products to investigate core research questions and deliver AI features that address practical needs.

Join Integrated Resources, Inc. as a Research Scientist, where you will play a crucial role in advancing scientific knowledge and innovation. You will engage in cutting-edge research projects, collaborate with interdisciplinary teams, and contribute to the development of impactful solutions in various scientific fields. Your expertise will help us drive our mission of delivering high-quality research that enhances lives.

Join our dynamic team at Anthropic as a Research Engineer or Research Scientist specializing in Tokens. In this role, you will engage in groundbreaking research and development, contributing to the advancement of our token systems and enhancing their capabilities. Your work will be pivotal in shaping the future of our technology and driving innovation in our projects.