ML Researcher​/Engineer - Voice and Speech to Speech Models; BM

Position: ML Researcher / Engineer - Voice and Speech to Speech Models (BM)
Location: Bengaluru

Job Description

About Apna:

Apna is India’s largest jobs and professional networking platform and one of India’s fastest unicorns. As we expand Blue machines AI, our AI-first platform across voice, text, and multimodal workflows — we’re looking for a bold and curious Applied ML Scientist who wants to shape the future of applied Gen AI.

Location:

Bengaluru (Work from Office - Domlur)
Team:  AI & Machine Learning

Experience:

2 –7 years
Requirements

What You'll do:

Fine-tune and deploy  LLMs, TTS, STT, and voice models  for use in real-time conversations with millions of users.
Convert  unstructured, messy real-world audio/text data  into clean, high-quality datasets for training and evaluation.
Build  inference pipelines  optimized for low-latency, high-accuracy voice agents and multimodal interfaces.
Work closely with infra and product teams to ship  production-grade GenAI models  with observability, fallback, and monitoring.
Experiment with  GANs, diffusion models, audio generation , and multimodal fusion to power next-gen AI agents.
Own the full model lifecycle —  from research and training to deployment, testing, and iteration.

What we're Looking for:

2-7 years of hands-on experience in AI / ML roles, ideally in  startups or product-driven teams.
Strong grasp of  LLM fine-tuning , instruction tuning, or pretraining techniques.
Familiarity with  TTS/STT systems , Whisper, Tacotron, VITS, or other open source models .

Experience with  multimodal architectures , generative audio, GANs, or diffusion-based models.
Ability to  work with real-world messy data , design training pipelines, and debug model failure modes.
Fluency in frameworks like  PyTorch, Hugging Face, Tensor Flow , and ecosystem tools (ONNX, Triton, Lang Chain, etc.).
Passion for building high-impact AI features that ship to real customers.

Why Join Us:

Work at the cutting edge  of LLMs, voice AI, and generative models — and ship real products, not just prototypes.
Directly impact millions of users  by powering AI agents that help with hiring, learning, and career growth.
Collaborate with a world-class team  of AI engineers, researchers, and product minds who move fast and ship boldly.
Freedom to explore:  Own experiments, propose architecture, or contribute to foundational model training.
Startup speed, enterprise scale —  best of both worlds. Rapid iteration and direct customer feedback.
Multilingual India - first problems  that push the boundaries of speech, reasoning, and personalization.