Speech Data scientist

This role is for one of the Weekday's clients

Salary range: Rs 2400000 - Rs 3500000 (ie INR 24-35 LPA)

Min Experience: 6 years

Location: Bangalore

JobType: full-time

We are seeking a skilled Speech Data Scientist to design, develop, and optimize advanced speech analytics and automatic speech recognition (ASR) solutions. The ideal candidate will work on end-to-end speech pipelines, multilingual audio processing, and model deployment in production environments. You will also drive research and innovation in speech processing, contributing to model enhancement and high-impact technical solutions.

Key Responsibilities

Core Development & Implementation

• Design and implement end-to-end speech analytics pipelines for production.
• Develop ASR engines using frameworks such as Wav2vec, Whisper, and Deep Speech with PyTorch or TensorFlow.
• Build and optimize speaker diarization, language identification (LID), and text post-processing systems.
• Focus on multilingual audio processing and domain adaptation strategies.
• Lead data selection and preprocessing for improved model performance.

Model Development & Enhancement

• Develop and analyze objective measures for speech quality evaluation and enhancement.
• Implement speaker-conditioned personalization techniques to improve ASR accuracy in noisy environments.
• Optimize on-device ASR models, emphasizing multi-language scenarios.
• Guide teams on best practices for model accuracy and performance optimization.

Research & Innovation

• Conduct research on advanced speech processing and neural speech enhancement techniques.
• Develop novel solutions for multi-speaker and complex audio scenarios.
• Contribute to patents, publications, and technical thought leadership in speech technology.
• Stay updated on transformer models, attention mechanisms, and foundation models.

Technical Integration & Deployment

• Design integration architectures for speech-to-text services and related technologies.
• Implement MLOps processes and CI/CD pipelines for speech model deployment.
• Deploy and scale speech solutions on cloud platforms (AWS, GCP).
• Develop production-ready applications using Python, C++, and Java.

Required Qualifications

Education

• Ph.D./M.S./M.Tech in Computer Science, Signal Processing, or related field preferred.
• B.Tech/B.E in ECE, CSE, or related technical field required.

Technical Expertise

• Speech Processing: 3–6 years of hands-on experience in ASR and speech analytics. Strong knowledge of HMMs, GMMs, ANNs, language modeling, CNNs, RNNs, LSTMs, CTC, and attention mechanisms.
• Machine Learning / Deep Learning: Proficiency in PyTorch and TensorFlow; experience with transformer models (BERT, Wav2vec 2.0, Whisper) and end-to-end ASR implementation.
• Programming & Tools: Strong Python skills (numpy, pandas, scikit-learn), experience with C++/Java for production, bash scripting, and Git.
• Cloud & Deployment: Hands-on experience with AWS/GCP, containerization (Docker, Kubernetes), MLOps, CI/CD pipelines, and scalable model serving.

Skills

ASR, Speech Recognition, Speech Analytics, Multilingual Audio Processing, Python, PyTorch, TensorFlow, Deep Learning