Principal Scientist, Machine Learning - Origination

ABOUT PIONEERING INTELLIGENCEPioneering Intelligence extends Flagship Pioneering's tradition of launching innovative scientific and computational initiatives. By leveraging the latest advancements in AI, machine learning, and data analytics, we aim to expedite fundamental research and build a diverse portfolio of AI-first companies. As a key player in Flagship's integrated approach to science, entrepreneurship, and investment, we transform groundbreaking concepts into transformative enterprises that enhance AI innovations in human health, sustainability, and more.THE ROLEWe are on the lookout for a Principal Scientist specializing in Embedded Machine Learning and Computational Techniques to spearhead various AI/ML and computational initiatives across early-stage ventures in our company origination process. You will be responsible for defining and executing practical AI strategies, overseeing the development of methodologies and platforms such as systems design, drug design, molecular modeling, systems biology, protein design, and LLM-driven workflows while ensuring the utmost rigor in model development, benchmarking, scaling, and reporting. You will manage cross-functional contributors as necessary, influence the strategic direction of the company, and represent Pioneering Intelligence to venture teams and external partners. The ideal candidate is a proactive, deep thinker who can seamlessly transition between protein design one week and mass spectrometry or docking pipelines the next, while also developing LLM-based agents to automate scientific workflows.

ABOUT PIONEERING INTELLIGENCE Pioneering Intelligence is a forward-thinking initiative that builds upon Flagship Pioneering's rich history of establishing groundbreaking scientific and computational enterprises. By leveraging cutting-edge advancements in artificial intelligence, machine learning, and data science, we aim to accelerate fundamental research and cultivate a dynamic portfolio of AI-first companies. As an integral part of Flagship's unique model that intertwines science, entrepreneurship, and investment, we transform revolutionary concepts into impactful companies, enhancing AI innovations that contribute to human health, sustainability, and more. THE ROLE We are on the lookout for a Principal Scientist specializing in Embedded Machine Learning and Computational Methods to spearhead various AI/ML and computational initiatives across early-stage ventures within our company origination framework. Your responsibilities will include defining and executing practical AI strategies, overseeing the development of methodologies and platforms (including systems design, drug design, molecular modeling, systems biology, protein design, and LLM-based workflows), and ensuring a high standard of rigor in model development, benchmarking, scaling, and reporting. You will also manage cross-functional teams as necessary, influence the strategic direction of our initiatives, and represent Pioneering Intelligence to venture teams and external collaborators. The ideal candidate is a self-motivated individual with a diverse skill set, capable of transitioning seamlessly from protein design to mass spectrometry or docking pipelines, and then creating LLM-based agents to streamline scientific workflows.

Are you passionate about advancing the field of machine learning? Join our team at Altos Labs as a Machine Learning Scientist or Senior Machine Learning Scientist. In this role, you will leverage your expertise to drive innovation and development in cutting-edge research projects. Collaborate with a multidisciplinary team to push the boundaries of machine learning technology.

Innovate and Create Models from Scratch! At Flagship Pioneering, we are dedicated to launching pioneering companies that challenge the status quo. Within our Flagship Labs, small, agile teams develop new technical theses, rigorously test them, and build ventures around groundbreaking concepts. We are assembling an exceptional machine learning team within our newly established venture, Flagship Labs 120. Our focus is on uncovering hidden structures from complex measurement data of intricate physical systems, often necessitating mechanism-informed modeling, carefully designed inductive biases, and principled methodologies for inverse problems. This is a pioneering role dedicated to innovative modeling, moving beyond standard optimization tasks. You will have the opportunity to design, prototype, test, and refine novel approaches that establish the technical foundation of our platform starting from day one.

Our MissionAt Altos Labs, our mission is to rejuvenate cell health and resilience, aiming to reverse diseases, injuries, and the associated disabilities that arise throughout life.For further details, visit our website at altoslabs.com.Our ValuesWe embrace a singular value at Altos: Everyone Owns Achieving Our Inspiring Mission.Diversity at AltosWe hold a firm belief that diverse perspectives are crucial to scientific advancement and inquiry. At Altos, outstanding scientists and industry leaders collaborate from across the globe to further our shared mission. We prioritize creating a sense of belonging, ensuring that every employee feels valued for their unique contributions. We are collectively responsible for maintaining a diverse and inclusive workplace.Your Contributions to AltosIn the role of Senior or Principal Machine Learning Scientist, you will be instrumental in the development of generative AI/ML models for multi-modal, multiscale biology, spanning from virtual cells to agentic target assessment. We seek a hands-on, innovative, and collaborative individual to join our multidisciplinary team of scientists and engineers dedicated to transforming the treatment of aging and disease. The ideal candidate will flourish in a dynamic environment that values teamwork, transparency, scientific excellence, originality, rigor, and integrity.

Join us at the Toyota Research Institute (TRI) as we strive to enhance the quality of human life through innovative technologies. Our mission is to create groundbreaking tools that enrich human experiences. To spearhead this transformative evolution in mobility, we've assembled a stellar team that is pushing the boundaries in artificial intelligence, robotics, driving, and material sciences.The TeamWithin TRI's Energy and Materials division, the Future Factory team is dedicated to pioneering advanced tools and methodologies that drive flexibility and efficiency in Toyota's product design and manufacturing processes. Our goal is to expedite the journey towards an emissions-free future. We are developing comprehensive AI systems capable of reasoning through the creation of physical objects — from initial design concepts to the assembly of actual components — and building the necessary infrastructure to train and evaluate these systems on a large scale.The OpportunityWe are seeking a Research Scientist to help us build intelligent systems for physical assembly. This role is an excellent fit for recent PhD graduates with a proven track record in implementation and a deep curiosity about the manufacturing process.As a member of our research team, you will design and implement learning pipelines from the ground up, conduct experiments to assess various architectural, data, and algorithmic alternatives, and influence the application of modern machine learning to the challenges of robotic assembly. Your work will intersect policy learning, reinforcement learning, and physical reasoning, allowing you to explore the integration of large language models and agentic infrastructure in solving real-world manufacturing challenges.

Your Impact at Lila As a Machine Learning Scientist specializing in Scientific Reasoning, you will be at the forefront of advancing AI systems designed to think like scientists. Your role involves crafting innovative frameworks that extend the capabilities of LLM-based reasoning methods while also developing scalable systems that enhance Lila’s platforms. This position merges profound theoretical insights with hands-on ML engineering, facilitating breakthroughs in the generation, testing, deployment, and optimization of scientific hypotheses. What You Will Be Creating Develop and formalize frameworks for scientific reasoning using LLMs, incorporating structured prompting, reasoning chains, and computational strategies at test time. Investigate and implement techniques for in-context learning, self-reflection, and adaptive reasoning within scientific discovery processes. Create scalable model prototypes aimed at addressing cutting-edge scientific challenges. Collaborate with scientists and engineers to embed domain knowledge into reasoning systems that unify symbolic and statistical methodologies.

Your Contribution at Lila Sciences As a Robotics Scientist at Lila Sciences, you will spearhead the research and innovation of autonomous robotic systems that form the intelligent physical backbone of our cutting-edge scientific superintelligence platform. You will be instrumental in developing groundbreaking algorithms and deploying advanced robotic solutions that engage seamlessly with human scientists and intricate laboratory settings. Your efforts will propel our mission forward by establishing fully autonomous workflows for scientific exploration, merging state-of-the-art robotics, machine learning, and systems engineering. Projects You Will Undertake Innovating advanced methodologies for precise and dexterous robotic manipulation that utilize foundation models, reinforcement learning, diffusion-based techniques, and human input, enabling adaptable and intelligent robotic systems capable of executing complex tasks in a variety of scientific environments. Creating novel frameworks for human-robot interaction that integrate imitation learning and learning from human feedback, demonstrations, and corrections, fostering intelligent robotic agents that can smoothly blend into human scientific workflows and swiftly adapt to new experimental scenarios. Enhancing dexterous manipulation research through avant-garde machine learning techniques, including diffusion models and adaptive learning algorithms, synthesizing multi-modal sensing (tactile, visual, and linguistic) to develop generative skill representations and sophisticated motor learning policies for intelligent robotic systems. Designing autonomous robotic systems equipped with trust calibration mechanisms, allowing intelligent agents to dynamically modify their behaviors based on contextual information in challenging scientific tasks.

Location: Cambridge, MA (Eastern Time / UTC -4). Relocation support is available. Remote work may be considered for candidates based outside Massachusetts.Start date: ASAPLanguages: English (required) Pragmatike is an AI startup founded by MIT CSAIL researchers, recognized by GTM Capital as a Top 10 GenAI company. The team develops advanced AI systems with a focus on real-world applications. Role overview The Principal Machine Learning Operations Engineer shapes the architecture and scaling of Pragmatike’s machine learning infrastructure. This senior position leads the design, implementation, and optimization of production AI systems, overseeing the full lifecycle from model training and evaluation to deployment, monitoring, and ongoing improvement. Collaboration is central in this role. The Principal ML Ops Engineer works closely with AI researchers, GPU systems engineers, backend developers, and product teams. The main focus is building reliable, efficient, and automated ML platforms to support large-scale AI deployments. Key responsibilities Architect, build, and improve the end-to-end ML Ops pipeline, including training, fine-tuning, evaluation, rollout, and monitoring. Design and maintain infrastructure for model deployment, version control, reproducibility, and orchestration across both cloud and on-premises GPU clusters. Optimize distributed systems for computational efficiency, covering Kubernetes, autoscaling, caching, GPU allocation, and checkpointing workflows. Establish observability for ML systems by monitoring model drift, performance, throughput, reliability, and operational costs. Automate workflows for dataset curation, labeling, feature engineering, evaluation, and CI/CD for ML models. Collaborate with researchers to bring models into production and improve training and inference pipelines. Set internal ML Ops standards, best practices, and develop tools that support cross-team collaboration. Mentor engineers and provide architectural guidance across the AI platform. Requirements Extensive hands-on experience designing and operating production ML systems at scale, ideally at the Staff or Principal level. Deep knowledge of ML Ops, distributed systems, and cloud infrastructure.

Your Impact at Lila Join our team as a Machine Learning Scientist focused on pioneering multi-modal reasoning through vision-language models (VLMs) leveraging real-world scientific data, including figures, plots, and microscopy data from various sources. Your innovative designs will contribute to the advancement of Scientific Superintelligence. What You Will Be Building Lead cutting-edge research on multi-modal reasoning systems that analyze scientific data (images, plots, text, etc.) using advanced and custom VLMs. Design and implement training, adaptation, and test-time strategies (e.g., instruction tuning, supervised learning, RLHF, RAG) tailored for scientific comprehension tasks. Create datasets and benchmarks from authentic scientific artifacts (e.g., microscopy images, spectra, protocols) to evaluate model performance. Develop perception modules (e.g., OCR, table/structure recognition, plot parsing) for handling multi-modal data types. Collaborate with domain scientists and engineers to transition research into production-ready systems for enhancing scientific superintelligence. What You’ll Need to Succeed A graduate degree in a relevant discipline (Computer Science/AI, Applied Mathematics/Statistics, Electrical Engineering) or a physical sciences field (Materials, Chemistry, Physics) with a strong focus on machine learning; or equivalent research/industry experience. A proven track record in multi-modal machine learning or VLMs, evidenced by deployed systems, publications, or contributions to open-source projects. In-depth understanding of scientific QA/benchmarks and custom evaluation design. Experience with multi-modal fine-tuning, document parsing, dataset curation, and benchmarking. Robust engineering skills utilizing modern machine learning frameworks (e.g., PyTorch, Hugging Face). Strong communication and collaboration skills in cross-functional environments. Bonus Points For Experience with scientific data modalities in laboratory settings, such as microscopy images. Publications in leading ML/CV/NLP conferences or demonstrable impact in applied industrial research. Contributions to open-source multi-modal tools, evaluation suites, or datasets. About Lila Lila Sciences stands at the forefront of scientific superintelligence, operating as the world’s first platform and autonomous laboratory dedicated to life sciences, chemistry, and materials science. We are committed to revolutionizing discovery by harnessing AI to enhance every aspect of the scientific method.

Join Cambridge Mobile Telematics (CMT), the largest telematics service provider globally, dedicated to enhancing road safety and improving driver behavior. Our AI-driven platform, DriveWell Fusion®, integrates sensor data from numerous IoT devices, including smartphones, connected vehicles, and dashcams, to provide actionable insights for auto insurers, automakers, and public safety organizations. With headquarters in Cambridge, MA, and offices across the globe, we are committed to making a difference for millions of drivers every day.As a Principal Machine Learning Engineer on the innovative DriveWell Atlas team, you will lead the charge in developing cutting-edge AI systems leveraging telematics data. You will spearhead projects that involve the design, pre-training, fine-tuning, and deployment of novel foundation models aimed at revolutionizing risk assessment, driver engagement, and claims processing. This role demands a profound understanding of contemporary AI methodologies, machine learning, and sensor data analysis, offering the chance to construct a pioneering large language model from petabyte-scale datasets.We seek a collaborative and inventive Principal Machine Learning Engineer who is driven by a passion for making roads safer and improving driver capabilities.

Join Cambridge Mobile Telematics (CMT), the largest telematics service provider globally, dedicated to enhancing road safety for all drivers. Our innovative AI-driven platform, DriveWell Fusion®, aggregates data from millions of IoT devices—including smartphones, proprietary Tags, connected vehicles, and dashcams—fusing it with contextual information to provide a comprehensive view of driver and vehicle behavior. Our insights are utilized by auto insurers, automobile manufacturers, commercial mobility companies, and public sector entities to enhance risk assessment, safety protocols, claims processing, and driver improvement initiatives. With our headquarters in Cambridge, MA, and offices in Budapest, Chennai, Seattle, Tokyo, and Zagreb, we safeguard millions of drivers worldwide every day.As a Principal Machine Learning Engineer at CMT, you will craft algorithms that model the physical phenomena associated with vehicle and human movements as captured by mobile sensors and IoT devices. Your expertise will help solve complex real-world challenges and drive innovative solutions through exploratory data analysis on intricate, high-dimensional datasets. Leveraging your knowledge in statistics, machine learning, deep learning, signal processing, physics, programming, and data modeling, you will spearhead Data Science projects, independently create new solutions to meet CMT’s goals, and contribute to company-wide prototype development and product enhancements.If you are a collaborative, customer-focused, and inventive Principal Machine Learning Engineer eager to contribute to safer roads and better driving practices, we welcome your application!

Collaboration Fuels Innovation. Join Roku in Transforming TelevisionAs the leading TV streaming platform in the U.S., Canada, and Mexico, Roku aims to power every television globally. We pioneered the streaming revolution, and our mission is to connect the entire TV ecosystem. We empower viewers with access to their favorite content, assist content creators in reaching vast audiences, and offer advertisers unparalleled engagement tools.From day one at Roku, your contributions will be recognized and valued. In our rapidly growing public company, you won’t just observe; you will play a vital role in enhancing the experience for millions of TV streamers across the globe while gaining invaluable experience across various fields. About the TeamThe Advanced Development team at Roku is at the forefront of innovation, creating the next generation of intelligent and generative media systems. We explore concepts that are years ahead of production, crafting foundational technologies that will redefine content understanding, creation, and personalization across millions of Roku devices.This unique environment is comprised of a PhD-level, interdisciplinary team that merges machine learning research, software engineering, and DevOps. Our experts possess not only deep technical knowledge but also a broad creative vision, challenging the status quo, embracing uncertainty, and building unprecedented solutions. Our culture is collaborative, low-ego, and driven by ownership, trust, and curiosity.We seek an Applied Scientist with a robust foundation in mathematics, machine learning, and computer science, complemented by experience in cloud engineering, DevOps, and computer vision — someone who excels at the intersection of research and production.

Your Contribution at Lila At Lila, we are assembling a dynamic and empowered AI safety team dedicated to proactively addressing the potential risks associated with scientific superintelligence. This team will collaborate closely with all core departments, including science, model training, and lab integration, to craft a customized safety strategy that aligns with our unique objectives and deployment methods. Key responsibilities will encompass the development of technical safety strategies, engagement with the broader ecosystem, and the creation of essential technical documentation, including risk assessments and capability evaluations. Your Key Responsibilities Design and execute evaluations to identify scientific risks—focusing on both established and emerging threats—from state-of-the-art scientific models integrated with automated physical laboratories. Develop initial proof-of-concept safety measures, such as machine learning models designed to detect and mitigate unsafe behaviors from scientific AI models and physical laboratory outputs. Gain a comprehensive understanding of various model capabilities, primarily within scientific contexts but also extending to non-scientific domains (e.g., persuasion, deception) to shape Lila's overarching safety strategy. Engage in high-quality research initiatives as needed to evaluate and restrict scientific capabilities effectively. Qualifications for Success A Bachelor's degree in a relevant technical field (e.g., computer science, engineering, machine learning, mathematics, physics, statistics) or equivalent experience. Proficient programming skills in Python and hands-on experience with machine learning frameworks (such as Inspect) for large-scale evaluations and structured testing. Demonstrated experience in constructing evaluations or conducting red-teaming exercises pertaining to CBRN/cyber risks or frontier model capabilities, encompassing both unsafe and benign attributes. Background in designing and/or implementing AI safety frameworks in cutting-edge AI enterprises. Exceptional ability to communicate intricate technical concepts and issues to audiences without technical expertise. Desirable Qualifications A Master’s or PhD in a field pertinent to safety evaluations of AI models within scientific areas or another technical discipline. Publications in AI safety, evaluations, or model behavior at leading ML/AI conferences (such as NeurIPS, ICML, ICLR, ACL) or model release documentation. Experience exploring risks arising from novel scientific advancements (e.g., biosecurity, computational biology) or utilizing specialized scientific tools (e.g., large-scale foundational models in science).

Your Impact at Lila As an ML Research Scientist specializing in Multimodal Data Extraction, you will play a pivotal role in advancing Lila's mission of achieving scientific superintelligence. Your work will focus on the development of foundational models capable of autonomously reading, interpreting, and organizing scientific knowledge from diverse formats such as text, images, and experimental data in the physical sciences. Your research will contribute to the unification of global scientific data into a machine-readable format, enhancing reasoning, prediction, and autonomous discovery within materials science and chemistry. What You Will Be Building Innovate and create AI systems that effectively extract and organize knowledge from a variety of scientific resources. Design and optimize large language models, multimodal models, and specialized architectures for accurate and interpretable data extraction. Develop scalable solutions for managing unstructured and heterogeneous scientific data, integrating various formats including text, tables, and visuals. Collaborate with subject matter experts to ensure that the extracted data aligns with real-world research workflows. Publish impactful research that propels the field of multimodal understanding and AI-driven knowledge extraction forward.

Join the Team at GraphcoreAt Graphcore, we are at the forefront of revolutionizing AI computation. Our team of experts in semiconductors, software, and artificial intelligence is committed to developing an advanced AI compute stack that encompasses everything from silicon to datacenter infrastructure.As a proud member of the SoftBank Group, we are supported by significant long-term investment, enabling us to drive crucial technology within the rapidly expanding SoftBank AI ecosystem.To address the immense opportunities within the AI landscape, we are seeking to expand our teams globally. We welcome the brightest minds to tackle the most challenging problems, ensuring that every team member has an opportunity to influence our company, our products, and the future of AI.

Imagine... being part of an innovative organization that cultivates and develops life sciences companies dedicated to pioneering groundbreaking technologies aimed at revolutionizing health care and sustainability. Company Overview Expedition Medicines is a dynamic, privately held, early-stage biotechnology company at the forefront of the Protein Editing field. We specialize in creating small molecules that modify protein structure and function to target currently undruggable options across various therapeutic strategies. Our advanced platform combines cutting-edge small molecule chemistry and chemoproteomic discovery technologies with Machine Learning (ML) to facilitate generative design. Backed by Flagship Pioneering, we leverage their vision, courage, and resources to propel Expedition Medicines from platform validation to impactful patient solutions. We seek passionate, collaborative, and determined problem solvers to join our mission! Role Overview: The Senior/Principal Scientist in Structural Biology will concentrate on determining protein-ligand complex structures using X-ray crystallography and cryo-electron microscopy (cryo-EM). This role involves delivering structural insights and collaborating with project teams to advance structure-based drug design and uncover the mechanisms of action of small molecules. Key Responsibilities: Design and implement structural biology strategies for multiple projects. Oversee the design, expression, and purification of protein constructs for structural determination. Develop and execute protocols to achieve high-quality crystals for X-ray crystallography. Create and implement protocols for grid optimization, data collection, and analysis in cryo-EM. Refine and communicate structures of protein-ligand complexes, collaborating with project teams to drive hit-to-lead and lead optimization processes. Manage contract research organizations (CROs) to ensure efficient delivery of protein expression and structural biology services. Enhance our capabilities and serve as a trusted authority in the field of structural biology.

Charles River Analytics, a proud member of the GRVTY family, is at the forefront of developing innovative solutions and advanced technologies designed to address some of the most challenging issues faced globally. Our dynamic team of tech entrepreneurs collaborates to enhance AI, robotics, intelligent sensing, and human-centered computing. The research and technology we develop not only propel government initiatives but also unveil new opportunities in the commercial sector. At Charles River, we take immense pride in cultivating a workplace that attracts and retains top-tier talent and creative problem-solvers in the industry. Join us as we strive to accelerate our mission-driven innovations!We are currently seeking a dedicated Senior/Principal Scientist with expertise in Human-Systems Integration, Human Factors, or Human-Computer Interaction. In this role, you will spearhead the development of cutting-edge tools and interfaces that empower users to effectively collaborate with technology in demanding environments.Your responsibilities will include leading and expanding teams of researchers, designers, and engineers to assess work domains and user needs, create novel human-AI interaction concepts, and evaluate interactive prototypes. You will also be responsible for defining and directing technical projects, while fostering robust relationships with clients, sponsors, and collaborators. As a subject matter expert in Human Systems, you will work alongside multidisciplinary teams to explore and pursue research opportunities across various technical domains.

Your Impact at Lila As a Machine Learning Engineer on the Physical Sciences team, you will play a pivotal role in developing and managing comprehensive, scalable machine learning workflows. These workflows will address a wide range of scientific challenges in materials science, chemistry, and physical sciences. Your contributions will be instrumental in advancing research initiatives focused on cutting-edge algorithms, driving towards the establishment of scientific superintelligence to tackle today’s most significant challenges in physical sciences. What You Will Build Design, implement, and sustain end-to-end ML pipelines, encompassing data ingestion, feature engineering, model training, evaluation, deployment, and monitoring. Productionize models and services while ensuring robust testing, observability, and documentation in collaboration with cross-functional software teams; develop CI/CD workflows and automated evaluations to facilitate safe and frequent releases. Work closely with domain scientists and platform engineers to translate research insights into high-performing, scalable systems. Participate in technical design reviews, establish coding standards, and mentor colleagues on best practices. What You’ll Need to Succeed BS, MS, or PhD in Computer Science, Engineering, or a related quantitative field, or equivalent industry experience. A solid foundation in Python software engineering, including testing, packaging, and typing; experience with machine learning frameworks such as PyTorch and Hugging Face. Experience deploying ML services in cloud-based environments (FastAPI/GRPC, containers, orchestration, cloud infrastructure). Hands-on experience with deploying models in production systems (LLMs, multimodal models, databases, RAG) along with strong debugging and profiling skills. Effective communication and collaboration in cross-functional settings. Bonus Points For Familiarity with scientific or engineering domains (materials, chemistry, physics) and related data formats/benchmarks. Experience in GPU optimization (CUDA, Triton, compilation, distributed training). Previous contributions to open-source ML or scientific software. Experience with workflow orchestration, data provenance, or large-scale computing environments. About Lila Lila Sciences stands as the pioneering platform for scientific superintelligence, offering an autonomous laboratory dedicated to life sciences, chemistry, and materials science. We are at the forefront of a new era of limitless discovery, harnessing AI to revolutionize research and innovation in these fields.