ML Engineer - Speech
Rezo.aiNoida, Uttar Pradesh, IndiaOn-Site
Needs a free Superjobs account. You'll finish on Rezo.ai's site.
Skill score
Summary
This ML engineer fine-tunes and adapts open-source speech models on proprietary call audio, builds training and evaluation pipelines for multilingual speech, and curates messy real-world audio data. The role suits an engineer with 3-4 years in ML, including at least 18 months on speech or audio, who can ship models to production.
- Function
- Data Science
- Estimated salary
- Est. ₹9.5–14 LPA
- Estimated for a mid-level role, ~3 yrs experience, India. Not published by the employer.
- Work mode and location
- On-Site, Noida, Uttar Pradesh, India
- Experience
- 3–4 yrs
- Employment type
- Full-time
Job description
Skills required
Top skills
- Python
- PyTorch
- Speech model fine-tuning
- Speech fundamentals
All skills (7)
- Audio data curation
- Messy real-world audio
- NeMo
- ESPnet
- LoRA/PEFT
- Distributed training
- Multilingual speech
Other details
Responsibilities: Fine-tune and adapt open-source speech models on our proprietary call audio Build training and evaluation pipelines for multilingual and code-mixed speech Own model quality against metrics that matter operationally — entity and numeric accuracy, latency to first response — not just aggregate error rates Design and run data curation at scale: pseudo-labelling, speech enhancement, quality filtering on messy real-world audio Work with our linguist on text normalisation and pronunciation handling Evaluate candidate architectures, make the call with evidence, and ship the result to production with the platform team Requirements Must have skills: 3–4 years in ML, with at least 18 months on speech or audio specifically Strong Python and PyTorch; comfortable reading a paper and implementing it Hands-on experience fine-tuning at least one production speech model Solid grasp of speech fundamentals — mel-spectrograms, acoustic models and vocoders, encoder-decoder vs transducer architectures, evaluation methodology, sampling rates and what they cost you Understanding of how modern speech systems are actually built: self-supervised encoders, neural audio codecs, LM-based generation, flow matching Experience with genuinely messy audio, not only clean benchmark datasets Nice to have: NeMo, ESPnet, SpeechBrain, or Coqui Telephony-band or contact-centre audio Multilingual or code-switched speech work LoRA/PEFT, distributed training Open-source contributions or publications in speech
Company info
About Rezo.ai
Rezo.ai provides Agentic AI and Decision Analytics for enterprises, using AI to turn customer conversations into measurable business results, with a delivery team accountable for the outcome. Its customers are enterprises, particularly contact centers seeking better customer experience performance.