NobleStack Private Ltd. is looking for a Speech AI Research Engineer only from Tricity(Chandigarh, Mohali, Panchkula). Please apply or call only if you are from Chandigarh, Mohali, Panchkula.
Experience:
- 2 to 3 years of experience in AI/ML.
About the Role:
We're building our own ASR, TTS, and voice cloning models for real-time VoIP applications. You'll train models from scratch, work with large speech datasets, and deploy them to production.
What You'll Do:
- Train and fine-tune ASR, TTS, and voice cloning models
- Build data pipelines: collection, cleaning, alignment, augmentation
- Optimize models for low-latency, real-time inference
- Evaluate models on WER, MOS, latency, and speaker similarity
Must have skills:
- Python
- 2+ years in ML / deep learning with Python and PyTorch
- Hands-on experience training speech models (ASR or TTS)
- Familiarity with toolkits like NeMo, ESPnet, SpeechBrain, or Coqui
- Experience with multi-GPU training and Hugging Face
- Linux
Nice to Have:
- Indic-language speech experience
- Voice cloning or neural vocoders (HiFi-GAN, BigVGAN)
- VoIP / telephony exposure (SIP, RTP, 8kHz audio)
- Model optimization (ONNX, TensorRT, quantization)
Location:
Job Types: Full-time, Permanent
Pay: ₹480,000.00 - ₹720,000.00 per year
Benefits:
Education:
Experience:
- Python: 1 year (Preferred)
- total work: 2 years (Preferred)
- AI/ML: 1 year (Preferred)
Work Location: In person
