Applied Scientist Intern (Audio)

Work set-up: 
Full Remote
Contract: 
Internships
Work from: 
New York (USA)

Offer summary

Qualifications:

Enrolled in a PhD program in deep learning, computer vision, speech/audio processing, or a related field., Published peer-reviewed papers in reputable AI research venues such as CVPR, ICLR, or Interspeech., Strong skills in exploratory data analysis techniques like dimensionality reduction and clustering., Experience with Multimodal Large Language Models (LLMs) and proficiency in Python and PyTorch..

Key responsibilities:

  • Investigate the composition of open source audio deepfake datasets.
  • Use Multimodal LLMs to create feature representations of audio deepfake data.
  • Identify data trends related to bias, fairness, and classification difficulty using feature representations.
  • Collaborate with scientists and engineers across the organization.

Reality Defender logo
Reality Defender Computer Hardware & Networking Startup https://realitydefender.com/
11 - 50 Employees
See all jobs

Job description

About Reality Defender

Reality Defender provides accurate, multi-modal AI-generated media detection solutions to enable enterprises and governments to identify and prevent fraud, disinformation, and harmful deepfakes in real time. A Y Combinator graduate, Comcast NBCUniversal LIFT Labs alumni, and backed by DCVC, Reality Defender is the first company to pioneer multi-modal and multi-model detection of AI-generated media. Our web app and platform-agnostic API built by our research-forward team ensures that our customers can swiftly and securely mitigate fraud and cybersecurity risks in real time with a frictionless, robust solution.

Youtube: Reality Defender Wins RSA Most Innovative Startup

Why we stand out:

  • Our best-in-class accuracy is derived from our sole, research-backed mission and use of multiple models per modality

  • We can detect AI-generated fraud and disinformation in near- or real time across all modalities including audio, video, image, and text.

  • Our platform is designed for ease of use, featuring a versatile API that integrates seamlessly with any system, an intuitive drag-and-drop web application for quick ad hoc analysis, and platform-agnostic real-time audio detection tailored for call center deployments.

  • We’re privacy first, ensuring the strongest standards of compliance and keeping customer data away from the training of our detection models.

Responsibilities

  • Investigate the composition of open source audio deepfake datasets

  • Use Multimodal LLMs to create rich feature representations of audio deepfake data

  • Use feature representations of audio deepfake datasets to identify data trends as they relate to measures of bias, fairness, and difficulty of classification between real and fake data.

  • Collaborate with scientists and engineers across the organization

About You

  • Currently enrolled in a PhD program in deep learning, computer vision, speech/audio processing, or a related field

  • Implemented and/or published peer-reviewed papers in reputable AI research venues such as CVPR, ICLR, Interspeech

  • Strong skills and intuition with exploratory data analysis techniques such as dimensionality reduction and clustering

  • Experience with Multimodal LLMs and an understanding of their architectures

  • Have 1+ years of programming experience in Python and model building in PyTorch; experience with audio models, e.g. HuBERT/wav2vec, would be a plus but not essential.

  • Team player with a positive attitude and good communication skills

Required profile

Experience

Industry :
Computer Hardware & Networking
Spoken language(s):
English
Check out the description to know which languages are mandatory.

Hard Skills

PyTorch (Machine Learning Library)Python (Programming Language)Deep LearningCluster AnalysisExploratory Data AnalysisDimensionality ReductionAudio Signal Processing

Other Skills

  • Teamwork
  • Communication
Are you interested?

Physicist Related jobs

See more Physicist jobs