Applied Scientist 2, Prime Video & Studios CoreTech

Amazon Prime Video is looking for passionate, talented, and inventive Applied Scientists with a strong machine learning background to help build industry-leading Audio, Video and Language technology using GenAI.
As part of the AI team in Amazon Prime Video, you will work on developing and fine-tuning LLMs and developing new large multimodal models to advance state-of-the-art in the translation, transcription and generation of multingual cinematic content. You will use the latest GenAI techniques to create and assess cinematic content like videos, audio, and captions, and more.

You will design novel approaches to integrate speech, vision, and text for holistic understanding and generation of content, and Collaborate cross-functionally with product managers, engineers, and domain experts to define and deliver high-impact solutions.

Your work will directly impact millions of our customers in the form of products and services that make use of speech and language technology on Prime Video content.

Key job responsibilities
We are looking to hire a top notch scientist with strong machine learning background, sharp research acumen and a hunger for solving ambiguous and challenging science problems. This person will work on developing and fine-tuning LLMs and developing new large multimodal models to advance state-of-the-art in the translation, transcription and generation of multingual cinematic content. They will be required to use the latest GenAI advancements to create and assess cinematic content like videos, audio, and captions, and more. Flavor of problems we are trying to solve: a) generating high quality subtitles in many languages meeting high customer satisfaction bar, b) developing foundation model capable of enabling many down stream tasks in localization and accessibility space, c) generating captions with plot pertinency, character identification etc. d) developing solution for assessing quality of subtitles, captions and metadata unified across different content types, linguistic vs non-linguistic aspects, different languages, and which mimics end user satisfaction while at the same time is explainable.

About the team
By 2027, Prime Video (PV) aspires to be the primary destination for customers to watch and the primary channel for partners to distribute premium cinematic content. The Content Reasoning, Enrichment & Localization (CoREL) science team’s mission is to develop AI to deeply understand the diverse facets of Prime Video's content across multiple languages, and to create content that power immersive, cinematic experiences. To drive this mission and delight our customers, the CoREL Science Bangalore team addresses the localization and accessibility aspects by a) understanding fine-grained information contained in content to enable generation of localized assets in different languages, b) generating subtitles and accessibility assets like captions across different content types and languages, c) assessing quality of generated (1P) and partner submitted (3P) assets and metadata to ensure defect free customer experience on PV.

Basic qualifications

- 3+ years of building models for business application experience
- Experience in patents or publications at top-tier peer-reviewed conferences or journals
- Experience programming in Java, C++, Python or related language
- Experience in building speech recognition, machine translation and natural language processing systems (e.g., commercial speech products or government speech projects)
- PhD, or Master's degree and 4+ years of CS, CE, ML or related field experience
- Other basic qualifications:
- a/ Proficient in Python, DL frameworks and developing ML models
- b/ Deep understanding of language models and LLMs and ability to adapt them to specific use cases (training/fine-tuning LLMs/foundation models)
- c/ Experience with audio and text is preferred,
- d/ Proven research expertise in solving complex and ambiguous problems
- e/ Ability to dive deep into model/algorithm performance and failure modes to push the limits of performance
- f/ Ability to review latest research and advancements in industry and academia and leverage them to solve the problems in our domain
- g/ Science rigor in experiment design to tackle ambiguity and several moving parts and ability to defend results with solid data analysis
- h/ Strong communication (writing) skills to convey the research outcomes in weekly/monthly updates

Preferred qualifications

- Experience using Unix/Linux
- Experience in professional software development
- Experience with localization preferred

Our inclusive culture empowers Amazonians to deliver the best results for our customers. If you have a disability and need a workplace accommodation or adjustment during the application and hiring process, including support for the interview or onboarding process, please visit https://amazon.jobs/content/en/how-we-hire/accommodations for more information. If the country/region you’re applying in isn’t listed, please contact your Recruiting Partner.