Captions is the AI-powered creative studio. Millions of creators around the world have used Captions to make their video content stand out from the pack and we're on a mission to empower the next billion.

Based in NYC, we are a team of ambitious, experienced, and devoted engineers, designers, and marketers. You'll be joining an early team where you'll have an outsized impact on both the product and company's culture.

We’re very fortunate to have some the best investors and entrepreneurs backing us, including Kleiner Perkins, Sequoia Capital, Andreessen Horowitz, Uncommon Projects, Kevin Systrom, Mike Krieger, Antoine Martin, Julie Zhuo, Ben Rubin, Jaren Glover, SVAngel, 20VC, Ludlow Ventures, Chapter One, Lenny Rachitsky, and more.

Check out our latest milestone and our recent feature on the TODAY show and the New York Times.

** Please note that all of our roles will require you to be in-person at our NYC HQ (located in Union Square) **

Responsibilities:

Conduct research and develop models to advance the state-of-the-art in generative video technologies, focusing on areas such as video in-painting, super resolution, text-to-video conversion, background removal, and neural background rendering.

Design and develop advanced neural network models tailored for generative video applications, exploring innovative techniques to manipulate and enhance video content for storytelling purposes.

Explore new areas and techniques to enhance video storytelling, including research into novel generative approaches and their applications in video production and editing.

Create tools and systems that leverage machine learning, artificial intelligence, and computational techniques to generate, manipulate, and enhance video content, with a focus on usability and scalability.

Preferred Qualifications:

PhD in computer science or related field or 3+ years of industry experience.

Publication Record: Highly relevant publication history, with a focus on generative video techniques and applications. Ideal candidates will have served as the primary author on these publications.

Video Processing Skills: Strong understanding of video processing techniques, including video compression, motion estimation, and object tracking, with the ability to apply these techniques in generative video applications.

Expertise in Deep Learning: Proficiency in deep learning frameworks such as TensorFlow, PyTorch, or similar, with hands-on experience in designing, training, and deploying neural networks for video-related tasks.

Strong understanding of Computer Science fundamentals (algorithms and data structures).

Benefits: Comprehensive medical, dental, and vision plans

Anything you need to do your best work

We’ve done team off-sites to places like Paris, London, Park City, Los Angeles, Upstate NY, and Nashville with more planned in the future.

Captions provides equal employment opportunities to all employees and applicants for employment and prohibits discrimination and harassment of any type without regard to race, color, religion, age, sex, national origin, disability status, genetics, protected veteran status, sexual orientation, gender identity or expression, or any other characteristic protected by federal, state or local laws.

Please note benefits apply to full time employees only.

Redmond, Washington, United States

Overview We are seeking a Principal Research Engineer to join our organization and help improve steerability and control Large Language Models (LLMs) and other AI systems. Our team currently develops Guidance, a fully open-source project that enables developers to control language models more precisely and efficiently with constrained decoding.

As a Principal Research Engineer, you will play a crucial role in advancing the frontier of constrained decoding and imagining new application programming interface (APIs) for language models. If you’re excited about links between formal grammars and generative AI, deeply understanding and optimizing LLM inference, enabling more responsible AI without finetuning and RLHF, and/or exploring fundamental changes to the “text-in, text-out” API, we’d love to hear from you. Our team offers a vibrant environment for cutting-edge, multidisciplinary research. We have a long track record of open-source code and open publication policies, and you’ll have the opportunity to collaborate with world-leading experts across Microsoft and top academic institutions across the world.

Microsoft’s mission is to empower every person and every organization on the planet to achieve more. As employees we come together with a growth mindset, innovate to empower others, and collaborate to realize our shared goals. Each day we build on our values of respect, integrity, and accountability to create a culture of inclusion where everyone can thrive at work and beyond. In alignment with our Microsoft values, we are committed to cultivating an inclusive work environment for all employees to positively impact our culture every day.

Responsibilities Develop and implement new constrained decoding research techniques for increasing LLM inference quality and/or efficiency. Example areas of interest include speculative execution, new decoding strategies (e.g. extensions to beam search), “classifier in the loop” decoding for responsible AI, improving AI planning, and explorations of attention-masking based constraints. Re-imagine the use and construction of context-free grammars (CFG) and beyond to fit Generative AI. Examples of improvements here include better tools for constructing formal grammars, extensions to Earley parsing, and efficient batch processing for constrained generation. Consideration of how these techniques are presented to developers – who may not be well versed in grammars and constrained generation -- in an intuitive, idiomatic programming syntax is also top of mind. Design principled evaluation frameworks and benchmarks for measuring the effects of constrained decoding on a model. Some areas of interest to study carefully include efficiency (token throughput and latency), generation quality, and impacts of constrained decoding on AI safety. Publish your research in top AI conferences and contribute your research advances to the guidance open-source project. Other

Embody our Culture and Values

Inria (Grenoble), France

human-robot interaction, machine learning, computer vision, representation learning

We are looking for highly motivated students joining our team at INRIA. This project will take place in close collaboration between Inria team THOTH and the multidisciplinary institute in artificial intelligence (MIAI) in Grenoble

Topic: Human-robot systems are challenging because the actions of one agent can significantly influence the actions of others. Therefore, anticipating the partner's actions is crucial. By inferring beliefs, intentions, and desires, we can develop cooperative robots that learn to assist humans or other robots effectively. In this project we are in particular interested in estimating human intentions to enable collaborative tasks between humans and robots such as human-to-robot and robot-to-human handovers.

Contact pia.bideau@inria.fr The thesis will be jointly supervised by Pia Bideau (THOTH), Karteek Alahari (THOTH) and Xavier Alameda Pineda (RobotLearn).

Location Sunnyvale, CA Bellevue, WA

Description Are you a passionate scientist in the computer vision area who is aspired to apply your skills to bring value to millions of customers? Here at Ring, we have a unique possibility to innovate and see how the results of our work improve the lives of millions of people and make neighborhoods safer. You will be part of a team committed to pushing the frontier of computer vision and machine learning technology to deliver the best experience for our neighbors. This is a great opportunity for you to innovate in this space by developing highly optimized algorithms that will work on scale. This position requires experience with developing Multi-modal LLMs and Vision Language Models. You will collaborate with different Amazon teams to make informed decisions on the best practices in machine learning to build highly-optimized integrated hardware and software platforms.

Excited to see you at CVPR! We’ll be at booth 1404. Come see us to talk more about roles.

Our team consists of people with diverse software and academic experiences. We work together towards one common goal: integrating the software, you'll help us build into hundreds of millions of vehicles.

As a Research Engineer, you will work collaboratively to improve our models and iterate on novel research directions, sometimes in just days. We're looking for talented engineers who would enjoy applying their skills to deeply complex and novel AI problems. Specifically, you will:

• Apply and extend the Helm proprietary algorithmic toolkit for unsupervised learning and perception problems at scale
• Carefully execute the development and maintenance of tools used for deep learning experiments designed to provide new functionality for customers or address relevant corner cases in the system as a whole
• Work closely with software and autonomous vehicle engineers to deploy algorithms on internal and customer vehicle platforms

B GARAGE was founded in 2017 by two PhD graduates from Stanford University. After having spent over five years researching robotics, computer vision, aeronautics, and drone autonomy, the co-founders set their minds on building a future where aerial robots would become an integral part of our daily lives without anyone necessarily piloting them. Together, our common goal is to redefine the user experience of drones and to expand the horizon for the use of drones.

The B GARAGE team is always looking for an enthusiastic, proactive, and collaborative Robotics and Automation Engineers to support the launch of intelligent aerial robots and autonomously sustainable ecosystems.

If you're interested in joining the B Garage team but don't see a role open that fits your background, apply to the general application and we'll reach out to discuss your career goals.

The Perception team at Zoox is responsible for developing the eyes and ears of our self driving car. Navigating safely and competently in the world requires us to detect, classify, track and understand several different attributes of all the objects around us that we might interact with, all in real time and with very high precision.

As a member of the Perception team at Zoox, you will be responsible for developing and improving state of the art machine learning techniques for doing everything from 2D/3D object detection, panoptic segmentation, tracking, to attribute classification. You will be working not just with our team of talented engineers and researchers in perception, but cross functionally with several teams including sensors, prediction and planning, and you will have access to the best sensor data in the world and an incredible infrastructure for testing and validating your algorithms.

We are looking for a Research Engineer, with passion for working on cutting edge problems that can help us create highly realistic, emotional and life-like synthetic humans through text-to-video.

Our aim is to make video content creation available for all - not only to studio production!

🧑🏼‍🔬 You will be someone who loves to code and build working systems. You are used to working in a fast-paced start-up environment. You will have experience with the software development life cycle, from ideation through implementation, to testing and release. You will also have extensive knowledge and experience in Computer Vision domain. You will also have experience within Generative AI space (GANs, Diffusion models and the like!).

👩‍💼 You will join a group of more than 50 Engineers in the R&D department and will have the opportunity to collaborate with multiple research teams across diverse areas, our R&D research is guided by our co-founders - Prof. Lourdes Agapito and Prof. Matthias Niessner and director of Science Prof. Vittorio Ferrari.

If you know and love DALL.E, MUSE, IMAGEN, MAKE-A-VIDEO, STABLE DIFFUSION and more - and you love large data, large compute and writing clean code, then we would love to talk to you.

Location Seattle, WA New York, NY

Description We are looking for an Applied Scientist to join our Seattle team. As an Applied Scientist, you are able to use a range of science methodologies to solve challenging business problems when the solution is unclear. Our team solves a broad range of problems ranging from natural knowledge understanding of third-party shoppable content, product and content recommendation to social media influencers and their audiences, determining optimal compensation for creators, and mitigating fraud. We generate deep semantic understanding of the photos, and videos in shoppable content created by our creators for efficient processing and appropriate placements for the best customer experience. For example, you may lead the development of reinforcement learning models such as MAB to rank content/product to be shown to influencers. To achieve this, a deep understanding of the quality and relevance of content must be established through ML models that provide those contexts for ranking.

In order to be successful in our team, you need a combination of business acumen, broad knowledge of statistics, deep understanding of ML algorithms, and an analytical mindset. You thrive in a collaborative environment, and are passionate about learning. Our team utilizes a variety of AWS tools such as SageMaker, S3, and EC2 with a variety of skillset in shallow and deep learning ML models, particularly in NLP and CV. You will bring knowledge in many of these domains along with your own specialties.

Gothenburg, Sweden

This fully-funded PhD position offers an opportunity to delve into the area of geometric deep learning within the broader landscape of machine learning and 3D computer vision. As a candidate, you'll have the chance to develop theoretical concepts and innovative methodologies while contributing to real-world imaging applications. Moreover, you will enjoy working in a diverse, collaborative, supportive and internationally recognized environment.

The PhD project centers on understanding and improving deep learning methods for 3D scene analysis and 3D generative diffusion models. We aim to explore new ways of encoding symmetries in deep learning models in order to scale up computations, a necessity for realizing truly 3D generative models for general scenes. We aim to explore the application of these models in key problems involving novel view synthesis and self-supervised learning.

If you are interested and present at CVPR, then feel free to reach out to Prof. Fredrik Kahl, head of the Computer Vision Group.

Vancouver

Who we are Established in 2017, Wayve is a leader in autonomous vehicle technology, driven by breakthroughs in Embodied AI. Our intelligent, mapless, and hardware-agnostic technologies empower vehicles to navigate complex environments effortlessly. Supported by prominent investors, Wayve is advancing the transition from assisted to fully automated driving, making transportation safer, more efficient, and universally accessible. Join our world-class, multinational team of engineers and researchers as we push the boundaries of frontier AI and autonomous driving, creating impactful technologies and products on a global scale

We are seeking an experienced researcher to be a founding member of our Vancouver team! We are prioritising someone with experience leading projects in AI applied to autonomous driving or similar robotics or decision-making domains, inclusive, but not limited to the following specific areas: Foundation models for robotics or embodied AI Model-free and model-based reinforcement learning Offline reinforcement learning Large language models Planning with learned models, model predictive control and tree search Imitation learning, inverse reinforcement learning and causal inference Learned agent models: behavioural, oral and physical models of cars, people, and other dynamic agents

Challenges you will own You'll be working on some of the world's hardest problems, and able to attack them in new ways. You'll be a technical leader within our diverse, cross-disciplinary team, helping teach our robots how to drive safely and comfortably in complex real-world environments. This encompasses many aspects of research across perception, prediction, planning, and control, including:

Actively contributing to the Science’s technical leadership community, inclusive of proposing new projects, organising their work, and delivering substantial impact across Wayve. Leveraging our large, rich, and diverse sources of real-world driving data Architecting our models to best employ the latest advances in foundation models, transformers, world models, etc, evaluating and incorporating state-of-the-art techniques into our workflows Investigating learning algorithms to use (e.g. reinforcement learning, behavioural cloning) Leveraging simulation for controlled experimental insight, training data augmentation, and re-simulation Scaling models efficiently across data, model size, and compute, while maintaining efficient deployment on the car Collaborating with cross-functional, international teams to integrate research findings into scalable, production-level solutions Potentially contributing to academic publications for top-tier conferences like NeurIPS, CVPR, ICRA, ICLR, CoRL etc. working in a world-class team, contributing to the scientific community and establishing Wayve as a leader in the field

What you will bring to Wayve Proven track record of research in one or more of the topics above demonstrated through deployed applications or publications. Experience leading a research agenda aligned with larger organisation or company goals Strong programming skills in Python, with experience in deep learning frameworks such as PyTorch, numpy, pandas, etc. Experience bringing a machine learning research concept through the full ML development cycle Excellent problem-solving skills and the ability to work independently as well as in a team environment. Demonstrated ability to work collaboratively in a fast-paced, innovative, interdisciplinary team environment. Experience bringing an ML research concept through to production and at scale PhD in Computer Science, Computer Engineering, or a related field

What we offer you The chance to be part of a truly mission driven organisation and an opportunity to shape the future of autonomous driving. Unlike our competitors, Wayve is still relatively small and nimble, giving you the chance to make a huge impact

Location Seattle, WA

Description Amazon's Compliance Shared Services (CoSS) is looking for a smart, energetic, and creative Sr Applied Scientist to extend and invent state-of-the-art research in multi-modal architectures, large language models across federated and continuous learning paradigms spread across multiple systems to join the Applied Research Science team in Seattle. At Amazon, we are working to be the most customer-centric company on earth. Millions of customers trust us to ensure a safe shopping experience. This is an exciting and challenging position to deliver scientific innovations into production systems at Amazon-scale that increase automation accuracy and coverage, and extend and invent new research as a key author to deliver re-usable foundational capabilities for automation.

You will analyze and process large amounts of image, text and tabular data from product detail pages, combine them with additional external and internal sources of multi-modal data, evaluate state-of-the-art algorithms and frameworks, and develop new algorithms in federated and continuous learning modes that can be integrated and launched across multiple systems. You will partner with engineers and product managers across multiple Amazon teams to design new ML solutions implemented across worldwide Amazon stores for the entire Amazon product catalog.

Location San Diego

Description

At Qualcomm, we are transforming the automotive industry with our Snapdragon Digital Chassis and building the next generation software defined vehicle (SDV).

Snapdragon Ride is an integral pillar of our Snapdragon Digital Chassis, and since its launch it has gained momentum with a growing number of global automakers and Tier1 suppliers. Snapdragon Ride aims to address the complexity of autonomous driving and ADAS by leveraging its high-performance, power-efficient SoC, industry-leading artificial intelligence (AI) technologies and pioneering vision and drive policy stack to deliver a comprehensive, cost and energy efficient systems solution.

Enabling safe, comfortable, and affordable autonomous driving includes solving some of the most demanding and challenging technological problems. From centimeter-level localization to multimodal sensor perception, sensor fusion, behavior prediction, maneuver planning, and trajectory planning and control, each one of these functions introduces its own unique challenges to solve, verify, test, and deploy on the road.

We are looking for smart, innovative and motivated individuals with strong theory background in deep learning, advanced signal processing, probability & algorithms and good implementation skills in python/C++. Job responsibilities include design and development of novel algorithms for solving complex problems related to behavior prediction for autonomous driving, including trajectory and intention prediction. Develop novel deep learning models to predict trajectories for road users and optimize them to run-in real-time systems. Work closely with sensor fusion and planning team on defining requirements and KPIs. Work closely with test engineers to develop test plans for validating performance in simulations and real-world testing.

Minimum Qualifications: • Bachelor's degree in Computer Science, Electrical Engineering, Mechanical Engineering, or related field and 6+ years of Systems Engineering or related work experience. OR Master's degree in Computer Science, Electrical Engineering, Mechanical Engineering, or related field and 5+ years of Systems Engineering or related work experience. OR PhD in Computer Science, Electrical Engineering, Mechanical Engineering, or related field and 4+ years of Systems Engineering or related work experience.Preferred Qualifications: Ph.D + 2 years industry experience in behavior and trajectory prediction Proficient in variety of deep learning models like CNN, Transformer, RNN, LSTM, VAE, GraphCNN etc Experience working with NLP Deep Learning Networks Proficient in state of the art in machine learning tools (pytorch, tensor flow) 3+ years of experience with Programming Language such as C, C++, Python, etc. 3+ years Systems Engineering, or related work experience in the area of behavior and trajectory prediction. Experience working with, modifying, and creating advanced algorithms Analytical and scientific mindset, with the ability to solve complex problems. Experience in Autonomous driving, Robotics, XR/AR/VR Experience with robust software design for safety-critical systems Excellent written and verbal communication skills, ability to work with a cross-functional team

Location Multiple Locations

Description

Members of our team are part of a multi-disciplinary core research group within Qualcomm which spans software, hardware, and systems. Our members contribute technology deployed worldwide by partnering with our business teams across mobile, compute, automotive, cloud, and IOT. We also perform and publish state-of-the-art research on a wide range of topics in machine-learning, ranging from general theory to techniques that enable deployment on resource-constrained devices. Our research team has demonstrated first-in-the-world research and proof-of-concepts in areas such model efficiency, neural video codecs, video semantic segmentation, federated learning, and wireless RF sensing (https://www.qualcomm.com/ai-research), has won major research competitions such as the visual wake word challenge, and converted leading research into best-in-class user-friendly tools such as Qualcomm Innovation Center’s AI Model Efficiency Toolkit (https://github.com/quic/aimet). We recently demonstrated the feasibility of running a foundation model (Stable Diffusion) with >1 billion parameters on an Android phone under one second after performing our full-stack AI optimizations on the model.

Role responsibility can include both, applied and fundamental research in the field of machine learning with development focus in one or many of the following areas:

• Conducts fundamental machine learning research to create new models or new training methods in various technology areas, e.g. large language models, deep generative models (VAE, Normalizing-Flow, ARM, etc), Bayesian deep learning, equivariant CNNs, adversarial learning, diffusion models, active learning, Bayesian optimizations, unsupervised learning, and ML combinatorial optimization using tools like graph neural networks, learned message-passing heuristics, and reinforcement learning.

• Drives systems innovations for model efficiency advancement on device as well as in the cloud. This includes auto-ML methods (model-based, sampling based, back-propagation based) for model compression, quantization, architecture search, and kernel/graph compiler/scheduling with or without systems-hardware co-design.

• Performs advanced platform research to enable new machine learning compute paradigms, e.g., compute in memory, on-device learning/training, edge-cloud distributed/federated learning, causal and language-based reasoning.

• Creates new machine learning models for advanced use cases that achieve state-of-the-art performance and beyond. The use cases can broadly include computer vision, audio, speech, NLP, image, video, power management, wireless, graphics, and chip design

• Design, develop & test software for machine learning frameworks that optimize models to run efficiently on edge devices. Candidate is expected to have strong interest and deep passion on making leading-edge deep learning algorithms work on mobile/embedded platforms for the benefit of end users.

• Research, design, develop, enhance, and implement different components of machine learning compiler for HW Accelerators.

• Design, implement and train DL/RL algorithms in high-level languages/frameworks (PyTorch and TensorFlow).