Applied Scientist II (Audio)

📋 Description

• • Reality Defender is at the forefront of the cybersecurity revolution, dedicated to combating the growing threat of deepfakes and AI-generated media. As an award-winning company backed by prominent investors like DCVC, Illuminate Financial, Y Combinator, and industry giants such as IBM and Accenture, we empower enterprises and governments worldwide to identify and neutralize malicious content. Our patented multi-model approach provides robust detection capabilities across video, audio, imagery, and text, ensuring that generative AI is not weaponized for fraud, disinformation, or harmful purposes. Our API-first platform is designed for real-time detection, making it an indispensable tool for security teams and developers.
• • As an Applied Scientist II specializing in Audio, you will play a pivotal role in shaping the future of audio deepfake detection. Your primary focus will be on the practical application and refinement of state-of-the-art machine learning and deep learning models. This involves not just building models, but critically, tuning, optimizing, and deploying them into demanding, real-world client environments. You will be instrumental in ensuring these models perform reliably, robustly, and efficiently, even when faced with adversarial attacks, edge-case scenarios, and the inherent complexities of diverse audio inputs.
• • Your responsibilities will extend to deep-diving into model performance, particularly investigating failure cases that emerge in production. This investigative work will be crucial for building sophisticated, custom evaluation frameworks that accurately assess model behavior. Based on these insights, you will design and implement effective mitigation strategies to enhance model robustness and resilience against evolving threats. A significant part of your role will involve driving continuous model iteration, focusing on improving performance across a wide spectrum of real-world conditions. This includes addressing challenges posed by audio compression artifacts, various types of background noise, the unique characteristics of telephony systems, and the intricacies of streaming pipelines.
• • You will be a key technical voice within the company, responsible for clearly and effectively presenting your technical findings and detailed model performance insights to internal stakeholders. This communication is vital for informing strategic decisions and fostering cross-functional alignment. Furthermore, you will collaborate closely with our Product and Engineering teams. This interface is essential for developing a profound understanding of the production environment, its constraints, and its operational nuances. By integrating this knowledge, you will ensure that our performance assessments are relevant, comprehensive, and directly address the challenges faced by our clients.
• • This role demands a blend of theoretical knowledge and hands-on practical application. You will be expected to leverage your expertise in model building, training, and rigorous benchmarking to push the boundaries of what's possible in audio deepfake detection. Your work will directly contribute to the integrity and trustworthiness of digital communications, protecting individuals and organizations from sophisticated AI-driven manipulation. You will be part of a dynamic team that is passionate about leveraging cutting-edge AI for defensive cybersecurity applications, making a tangible impact in a rapidly evolving technological landscape.