Senior Voice AI / Speech ML Engineer

Location: HSR Layout, Bengaluru | Work Mode: On-site | Experience: 3–4 Years | Employment: Full-time
About the Role
We are looking for a Senior Voice AI / Speech ML Engineer to build and train next-generation speech intelligence systems. This role is for an engineer who has hands-on experience training speech models and foundation models, not just integrating existing APIs or pretrained models.
What You’ll Work On
Develop, train, fine-tune and evaluate speech/foundation models for Voice AI.
Build and improve ASR, TTS, Speaker ID, VAD and conversational AI systems.
Work with large-scale speech, audio and text datasets.
Design training pipelines, experiments and evaluation frameworks.
Optimize models for accuracy, latency, robustness and inference efficiency.
Work with PyTorch/TensorFlow, Transformers and GPU-based training.
Deploy and optimize models for production environments.
Work on multilingual and Indian-language speech AI use cases.
Must-Have
3–4 years of relevant post-graduation/full-time experience in ML, Speech AI, NLP or related fields.
Personally trained/pre-trained foundation models — not only fine-tuned or consumed existing models.
Hands-on experience developing/training ASR or TTS models.
Strong Python + PyTorch/TensorFlow.
Experience with Transformers, GPUs and large-scale model training.
Strong understanding of model evaluation, optimization and debugging.
Good to Have
Indian/multilingual speech experience.
Whisper, wav2vec 2.0, HuBERT, NeMo, SpeechBrain or Hugging Face.
Distributed training, quantization and production deployment.
Audio preprocessing, augmentation, annotation and dataset quality.
Not a Fit If
Only GenAI/RAG/LLM application development.
Only prompt engineering or API integration.
Only using pretrained Whisper/TTS models.
Only fine-tuning models without foundation-model training experience.
Less than 3 years relevant post-graduation/full-time experience.

Work arrangement
No