Senior ML Research Scientist

Blue Machines AI; Bengaluru; On-site; INR 4,000,000 - 8,000,000; Full time

Job summary

Blue Machines AI in Bengaluru is seeking an accomplished ML researcher to own a research direction in speech (STT, TTS, or speech-to-speech). You will design architectures, pre-train foundation models from scratch on multilingual audio, and mentor engineers.

In 12 months, you will ship a model pre-trained from scratch that beats the vendor it replaces on our benchmark, with a reproducible training and evaluation stack and a team of engineers producing results.

Qualifications

• 3–6 years in ML with at least 2–3 years focused on speech (ASR, TTS, or audio-language models).
• MS or PhD in CS, EE, or a related field, or equivalent depth shown through shipped work.
• Track record of pre-training at least one speech or audio model from scratch that reached production or strong published results.
• Deep knowledge of modern architectures: Conformer/Zipformer, RNN-T/TDT, encoder-decoder ASR, flow-matching and diffusion TTS, neural codecs (EnCodec, DAC, Mimi).

Responsibilities

• Lead one research track and set its roadmap with the CTO and product leads, tied to customer metrics (accuracy, naturalness, latency, cost per minute)
• Design novel architectures and pre-train speech foundation models from scratch on large multilingual corpora (100k+ hours), including self-supervised pre-training, tokenizer and codec design, and scaling experiments. Fine-tuning and continued training are tools, not the whole job.
• STT: streaming ASR with low time-to-first-token, code-switch robustness, contextual biasing for names and entities, and domain adaptation for BFSI and healthcare.
• TTS: expressive, low-latency streaming TTS for Indian languages; voice cloning, prosody and emotion control, and pronunciation of numbers, dates, and mixed-script text.
• Speech-to-speech: Full-duplex conversational models built in-house: our own audio tokenizers and codecs, speech-LLM pre-training and alignment, interruption and turn-taking, and end-of-utterance prediction.
• Build the data strategy: sourcing, licensing, synthetic data generation, labelling quality, and consent-compliant use of production audio.
• Define evaluation standards and benchmarks that predict real call outcomes, and run honest comparisons against external vendors.
• Make serving trade-offs with the platform team: distillation, quantization, GPU concurrency per model, on-prem footprints.
• Mentor 1-3 engineers; review experiment designs and code.
• Publish or open-source selectively where it strengthens our position.

Skills

Speech model pre-training

Research leadership

Experiment design

Distributed training

Clear communication

Education

MS or PhD in CS/EE or related field

Tools

Conformer

RNN-T

Transformer models

Codecs (EnCodec)

Job description

You will own a research direction (STT, TTS, or speech-to-speech) end-to-end. That means designing new architectures, pre-training foundation models from scratch on large multilingual audio, and getting them into production, while mentoring engineers on the team.

Responsibilities:

• Lead one research track and set its roadmap with the CTO and product leads, tied to customer metrics (accuracy, naturalness, latency, cost per minute)
• Design novel architectures and pre-train speech foundation models from scratch on large multilingual corpora (100k+ hours), including self-supervised pre-training, tokenizer and codec design, and scaling experiments. Fine-tuning and continued training are tools, not the whole job.
• STT: streaming ASR with low time-to-first-token, code-switch robustness, contextual biasing for names and entities, and domain adaptation for BFSI and healthcare.
• TTS: expressive, low-latency streaming TTS for Indian languages; voice cloning, prosody and emotion control, and pronunciation of numbers, dates, and mixed-script text.
• Speech-to-speech: Full-duplex conversational models built in-house: our own audio tokenizers and codecs, speech-LLM pre-training and alignment, interruption and turn-taking, and end-of-utterance prediction.
• Build the data strategy: sourcing, licensing, synthetic data generation, labelling quality, and consent-compliant use of production audio.
• Define evaluation standards and benchmarks that predict real call outcomes, and run honest comparisons against external vendors.
• Make serving trade-offs with the platform team: distillation, quantization, GPU concurrency per model, on-prem footprints.
• Mentor 1-3 engineers; review experiment designs and code.
• Publish or open-source selectively where it strengthens our position.

What success looks like in 12 months:

• Your track has shipped a model pre-trained from scratch by Blue Machines that beats the vendor it replaces on our benchmark and reduces cost or latency
• A clear, reproducible training and evaluation stack for your area
• Engineers you mentor are running experiments independently

Requirements:

• 3-6 years in ML, with at least 2-3 years focused on speech (ASR, TTS, speaker/voice, or audio-language models).
• MS or PhD in CS, EE, or a related field, or equivalent depth shown through shipped work.
• Track record of pre-training at least one speech or audio model from scratch (not only fine-tuning) that reached production or strong published results.
• Deep knowledge of modern architectures: Conformer/Zip former, RNN-T/TDT, encoder-decoder ASR, flow-matching and diffusion TTS, neural codecs (EnCodec, DAC, Mimi), and speech-LLMs.
• Large-scale pre-training experience: distributed training (FSDP/DeepSpeed/Megatron) on multi-node GPU clusters, 100k+ audio hours, scaling laws and compute budgeting, data loading at scale, and debugging loss spikes and unstable runs.
• Strong experimental design: ablations, significance, and avoiding test-set leakage.
• Clear written and spoken communication with engineers, products, and customers.

Nice to have:

• Publications at Interspeech, ICASSP, ACL/EMNLP, NeurIPS, or similar.
• Experience with Indic languages, low-resource languages, or code-switching.
• Built streaming or real-time speech systems with strict latency SLAs.
• Experience with RL or preference tuning for speech (e. g., naturalness rewards) or LLM post-training.
• Contributions to open-source speech toolkits.