Live Playback and Operations organization owns the live and 24/7 linear channel playback experience from content ingestion and signal acquisition, encoding (H.264/AVC, H.265/HEVC, AV1 etc) and streaming technologies (DASH, HLS, Smooth streaming, Low Latency), feature development (time shifted viewing, replays) to content distribution and delivery (CDN capacity management & load balancing) to the end customer for Prime Video. In addition, the organization handles the D2C (direct to consumer) app and Media Asset delivery for ~50 affiliates for MGM+. To run everything end-to-end at Amazon scale the org owns operations (active monitoring and incident response), instrumentation, infrastructure (on-premise & cloud encoders, direct connects, transport fiber, satellite), and tooling for these areas. Prime Video streams thousands of Amazon produced and third party live events per year in addition to thousands of linear stations globally. We achieve that with high-availability, low-latency, using different formats on approximately 1 billion devices, and do that for millions of customers with the best possible audio-video experience and frugal use of infrastructure and people. This is made possible through the use of advanced technology and science to automate operations, high judgment decision support systems and the infrastructure to deliver complex services simply, reliably and consistently. The Principal Applied Scientist leads innovation across a broad range of specialist areas. These include use of AI for routing content through our vast video delivery network, optimization of video operations, automation of signal processing, identification of defects and anomalies, predicting granular delivery capacity needs using public and private data, and much more. We have advanced capabilities in dedicated groups for low latency content delivery and computer vision quality analysis. Our organization is globally dispersed with an existing science group that has research collaborations with multiple US and European Universities, science publishing, contributions to industry standards and patents. The leader in this role will grow our science broadly across the organization, opening opportunities for innovation and use of cutting-edge science in areas where we see the potential to make significant impact to the service we provide to our customers and how we do that efficiently and effectively at Amazon scale.The scope of the role is broad and reports into the Director of the organization with the intention of steering our science strategy across a wide range of areas. This includes spotting opportunities, connecting with science and engineering leaders across Amazon. It is also required that the leader brings industry and university research context into the organization. The technology landscape changes fast and we make careful choices to balance vendor partnership along with self-build decisions to trade-off moving fast with owning core capabilities. The problems and challenges we face tomorrow may be quite different from those today, and the technology we need to innovate for our customers to go faster, increase our quality, and expand the range of products we provide to customers. The Principal is a key advisor to the Director and leadership team of the organization with the skills to communicate complex subjects simply both orally and in written form to influence our strategic and tactical decisions. You probe assumptions, illuminate pitfalls, bring clarity, and generate shared understanding on science topics. As an Applied Scientist, it is a 'hands-on' science leadership role, and requires opining on decisions that affect the way we create, build, and evaluate scientific solutions. Your views are influential in which research directions we invest in, the universities we collaborate with, and the outcome results. You grow other science leaders, mentoring and coaching them and develop science knowledge and application widely across the organization through educating leaders at multiple levels. You acquire expertise yourself to keep across trends and steer Prime Video in the best direction to invest and the benefits that will accrue from that. Your influence stretches across multiple, diverse teams, globally, and across organizational boundaries. Our science overall is stronger, our understanding of science topics is deeper and broader, and the programs we invest in generate impactful outcomes through your leadership.The role expects you to have a direct science output yourself in a strategic area, connecting with professors at leading universities, speaking and representing Amazon at conferences and industry events and influencing industry standard bodies. You are a thought leader and represent us on specific topics. You are connected with industry and standardization bodies and recognized in those communities. You publish, co-author, and review publications and secure our intellectual property through patents. Key job responsibilitiesThe Playback Development & Operations Principal Scientist is a subject matter expert recognized as a thought leader in a particular science discipline at a global level. For instance, you serve as an Area Chair at peer-reviewed conferences or as a senior program committee member. The role requires connection with internal and external research groups and you are expected to be active in the academic community and have relationships with multiple universities. Your track record includes research, publications, or intellectual property, and application of that domain expertise into real-world problems. The role expects that you will continue to publish in top-ranking journals and present at leading conferences. As a trusted advisor to senior leadership, you are an expert in state-of-the-art methodologies in at least one area of computer science, natural language processing, computer vision, video signal processing, Gen AI, content distribution, networking or other similar topics. As the senior science leader in a large organization, you will develop talent in Prime Video, through involvement in the hiring process, annual evaluations, and promotion assessments. You will develop other science leaders and have a background of growing multiple other experts to an advanced level, perhaps as a PhD advisor or as an industry leader. You have prior experience of getting technology across the line, and have written critical path code or algorithms in the past, and you could be hands on and own most complex and key algorithms when necessary. You understand the concerns of both software engineers and are familiar with research methods used in modern science. In current and previous roles, you have examples where you delivered scientifically complex solutions into production that are innovative and impactful. You can articulate those problems clearly, communicating core science concepts simply, and the resulting benefits. You are up to date with modern trends in compute science. You can speak confidently and influence opinion both at VP level and with junior engineers on the application and opportunities of Machine Learning and Generative AI. You are familiar and have hands-on experience using large language models, transformer architectures, and recent innovations. Ideally, you are also familiar with an aspect of audio-video signal processing, computer vision, or networking/media distribution technology and research. You will prepare and deliver artifacts that set the standard for Prime Video science leadership and engineering excellence. Your science writing and where applicable algorithms, analysis, and research results serve as example within the organization. The research you lead and influence will have results that are demonstrated in service/product features and efficiencies in operations. You collaborate effectively with other leaders to get things done. Your involvement makes science easier to generate impact and saves effort for the organization as a whole to progress our objectives. You do that through advising senior leaders and bringing understanding to complex subjects. You move fast, both to resolve ambiguity and bring innovative solutions to intrinsically hard, previously unsolved problems. To do that, you identify opportunities and convince other leaders by bringing them along with your ideas. You use metrics, data, and analysis to make informed decisions. You lead research that takes balanced risks that inform our future. You not only dive into the immediate pain points but also set the long term science roadmap and vision for the next 3-5 years.BASIC QUALIFICATIONS* PhD with applied research experience in Computer Science or in a highly quantitative field* Evidence of Science/Engineering publications in relevant journals or conferences with more than one as first author * Examples of sourcing funding / creating research relationships with universities or other institutions* Experience of building machine learning models for business application* Background of programming in Java, C++, Python or related language* Theoretical understanding of broad machine learning concepts, with deep and demonstrable expertise in at least one topic or application of machine learning* Conversant with current approaches to generative AI, large language models and their applications
...