Qualifications
Key ResponsibilitiesCreate and refine neural audio codecs for sound, music, and speech that advance sound quality and are tailored for generative model applications. Address the unique challenges of utilizing codecs as latent representations within generative audio frameworks, aiming to create top-tier audio generative models. Investigate the advantages and disadvantages of continuous audio representations (common in diffusion models) versus discrete representations (typically used in autoregressive models). Establish benchmarking pipelines for codec assessment. Conduct preliminary experiments with generative models to confirm the utility of new codec candidates for downstream applications. Ideal Candidate ProfileA strong background in deep learning for audio, including neural codecs, source separation, speech models, or generative audio systems. Hands-on experience in designing and training neural audio codecs. A solid grasp of audio signal processing principles. A proven track record in research and development in relevant fields.
About the job
Join Mirelo AI in revolutionizing creative tools by crafting realistic sound, speech, and music from video content.
We are at the forefront of developing innovative generative AI models that enhance silent video content, delivering custom, hyper-realistic audio for gaming, video platforms, and creators. Our technology empowers storytellers around the world to elevate their content.
Having recently secured a $41 million Seed funding round co-led by Andreessen Horowitz and Index Ventures, with contributions from Atlantic, we are rapidly expanding our teams across Product, Engineering, Go-to-Market, and Growth initiatives.
About the Role
As a Research Scientist at Mirelo, you will play a pivotal role in pushing the boundaries of generative audio. Your innovation will directly impact the quality of our underlying audio representations. You will engage with our modeling stack, designing, training, and assessing neural audio codecs that will influence the performance of our next-generation music and sound models. Collaboration with the model team will be essential as you explore both continuous and discrete representations and develop evaluation tools to measure our progress in building superior generative audio models.
About Mirelo AI
Mirelo AI is pioneering the future of creative tools through the generation of realistic audio elements from video, empowering creators worldwide to enhance their storytelling capabilities. With significant funding and a commitment to innovation, we are expanding rapidly and seek talented individuals to join our mission.
