Job DescriptionAI Engineer (Speech and Audio Systems)Job Summary:We are seeking a motivated, detail-oriented AI Engineer to join our growing Audio Intelligence team. In this role, you will help build, refine, and deploy machine learning models that understand human speech and sound. You will bridge the gap between creative audio engineering and deep learning by processing raw audio datasets, evaluating model performance, and helping us build the next generation of voice-activated technology.
Responsibilities:- Data Curation & Preprocessing: Clean, segment, and format audio datasets for model training (handling noise reduction, volume normalization, and formatting).
- Audio Analysis & Visual Inspection: Use advanced spectral editing software to inspect, clean, and isolate specific sound events or speech anomalies in our training data.
- Model Training Support: Assist in fine-tuning state-of-the-art automatic speech recognition (ASR) and audio classification models.
- Evaluation & Benchmarking: Run validation tests on trained models, track performance metrics (like Word Error Rate), and document failures to help guide future improvements.
- Collaboration: Work closely with senior AI scientists and data engineering teams to maintain high-quality data pipelines.
Qualifications:- Bachelor's degree in Computer Science, Data Science, Electrical/Audio Engineering, or a related technical field
- Understanding of basic digital audio concepts like sampling rates, bit depth, channels, etc.
- Familiarity with sound processing software such as SpectraLayers, Audacity, and python libraries such as scipy.signal, pydub, etc.
- Strong foundational programming skills, familiarity with version control Git and software engineering best practices
- Prior experience working in a fast-paced, high-growth environment (big tech or startup experience preferred).
Disclaimer: The above describes the general responsibilities, required knowledge and skills. Please keep in mind that other duties may be added or this description may be amended at any time.
#LI-SS2
