Senior Applied Scientist - Multimodal

"The AI company that's revolutionizing Hollywood" Flawless is transforming Hollywood with assistive AI. Our tools empower filmmakers to edit, localize, and refine performances while preserving artistic intent. Designed to support, not replace, artists, our technology expands what is possible on screen and gives creators freedom to tell stories with greater impact and reach audiences in new ways. From enabling seamless multilingual releases to eliminating the need for costly reshoots, Flawless solves critical challenges that slow down productions and limit distribution. We are also setting the standard for ethical AI in entertainment. Our Artistic Rights Treasury (A.R.T.) is a rights management solution that protects artists and rights holders, ensuring that innovation moves forward with transparency and respect for creative ownership. Reports to: Ari Shapiro What we are looking for: We’re looking for a deeply technical, product-driven applied scientist to help scale and operationalize our audio/video dataset generation, multimodal end-to-end pipelines, and lip sync work. This role exists to support and amplify ongoing research by owning model training pipelines, metrics, evaluation, and data workflows. This will be ensuring our audio/video, lip sync models improve reliably with every release. You’ll operate at the intersection of research and production, bringing rigor, automation, and clarity to how we validate and ship model improvements. Responsibilities: Model Development & Training • Develop repeatable, scalable audio/video dataset curation pipelines and lip sync model training workflows across multiple datasets • Train, fine-tune, and manage audio/video and lip sync model variants as model dependencies, data, and architectures evolve • Incorporate new datasets and model updates as they become available Evaluation, Metrics & Quality • Design, automate, and maintain audio/video datasets and lip sync metric testing pipelines • Generate new quantitative and qualitative metrics to evaluate audio/video and lip sync quality • Produce comparisons, visualizations, and analyses to inform research and product decisions Collaboration & Support • Partner closely with audio/video and lip sync researchers to support ongoing and future research initiatives • Validate audio/video and lip sync quality to improve out-of-the-box approval rates and reduce downstream cost and iteration time • Collaborate with Science, Engineering, and Product teams to align research outputs with company goals Qualifications: MSc OR PhD + Industry experience working in the domains of Audio processing, 3D Computer Vision, Speech Synthesis, Computer Graphics, or other multimodal related fields such as text/audio, or audio/visual. Proficiency in Python, with a strong foundation in computer science and problem-solving. Expertise with deep learning frameworks (PyTorch) and vision tools (OpenCV). A strong product mindset — motivated by building systems that deliver tangible value to users, not just technical novelty. Comfortable working at both the algorithmic and implementation levels, from model design and optimisation to large-scale data processing and integration in production systems. High degree of proficiency in math and statistical methods for signal processing Experience with audio-visual learning, multimodal fusion, and/or audio-driven face animation Experience with speech processing and detection, such as dialog/speaker detection, speaker separation, and speech synthesis with deep neural networks Outstanding communication skills for collaboration with scientists, research/ML engineers, and VFX artists Bonus points for: Demonstrable research experience with a strong publication record in major 3D Computer Vision, Speech Processing, and Computer Graphics venues and journals (e.g., CVPR, SIGGRAPH, NeurIPS) Experience developing multi-modal systems that integrate audio, text, and visual inputs. Experience working with cross-functional teams Experience with generative and cross-domain attention models for audio/visual-based speech applications Interview Process: At Flawless, our team and interview process want to help you show your best self. We’ll dive into past projects and simulate working together. Our interview process is three rounds with some casual Zoom (or in-person) coffee in between to get to know each other: - Recruiting Screen: 30-45 minute call with our recruiting team (We want to discuss your interests and motivations as well as the practical details and make sure that Flawless would be a good fit for you) - Hiring Manager Screen: 45-60 minute - Skills Interviews: A take home task to assess your coding ability and design decisions, this will be followed by a conversation to discuss your work and how it could be improved. - Team Interview: 2 hours onsite Interview where you will meet variety of your potential future colleagues. We will review your coding solution, discuss relevant papers and their application and have behavioural focussed round. Your Recruiter and hiring manager will be your main point of contact and prepare you for interviews. You’ll meet 4 to 6 people from across the business. (We make sure that you have time in each interview to ask them questions). If we don’t give an offer, we’ll provide feedback! Why work at Flawless? You will be working in an environment based on trust, autonomy and collaboration, and this is a great opportunity for someone who wants to be part of a growing company in its most exciting stage of development. You can play a part in shaping the future of a company that’s caring, creative and collaborative. In addition to this, you'll also receive: - Autonomy - A hybrid working environment - Competitive Salary - All permanent employees receive generous stock options I don’t meet all the listed requirements—should I still apply? Absolutely! Research shows that women and underrepresented groups often hesitate to apply unless they meet every qualification, but at Flawless, we actively work to break down those barriers. We believe diverse perspectives, experiences, and backgrounds make us stronger, and we are committed to supporting and elevating underrepresented talent. If you're excited about the role, share our values, and believe you can contribute meaningfully, we encourage you to apply—even if you don’t meet every single requirement. Your unique skills and perspective matter, and we’d love to hear from you ❤️