Machine Learning Engineer (Speech/Audio) - Singapore

PLAUD PTE. LTD.; Singapore; On-site; SGD 90,000 - 140,000; Full time

Job summary

Plaud Inc. is building the next generation of real-world AI interfaces for professionals. We are hiring for a role focused on speech and data engineering to enhance model training and deployment.

You will own end-to-end data pipelines, contribute to training and fine-tuning SpeechLLM or related models, and drive domain adaptation across languages and verticals with a strong emphasis on accuracy and robustness.

Qualifications

• Minimum 1 year hands-on experience in speech, ML, or large-scale data engineering.
• Experience in ASR/SpeechLLM or LLM/fine-tuning is a plus.
• Strong Python and PyTorch fundamentals.

Responsibilities

• Own large-scale speech/audio data pipelines for collection, cleaning, labeling, augmentation and quality control.
• Support model training and evaluation to improve recognition accuracy for domain terms.
• Perform domain adaptation to improve recognition across languages and verticals.
• Build test sets and evaluation frameworks and benchmark against baselines.

Skills

Speech/ML experience

Python

PyTorch

Distributed data processing

Data engineering

Tools

Python

PyTorch

Spark

Ray

Job description

About Plaud Inc.

Plaud is building the real-world AI interface for professionals to amplify intelligence, elevate productivity and performance, loved by over 2,000,000 users worldwide since 2023. With a mission to amplify human intelligence, Plaud captures, structures, and compounds the intelligence generated in conversations - so humans can think better, decide faster, and execute with clarity.

Plaud Inc. is a Delaware-incorporated, San Francisco-based company pushing the boundary of human-AI intelligence through a hardware-software combination. With full ISO 27001, ISO 27701, SOC 2, GDPR, EN18031, and HIPAA compliances, Plaud is committed to the highest standards of data security and privacy protection.

Why You Should Join Us

Plaud is building the next generation intelligence infrastructure and interfaces to capture, extract, and utilize intelligence from what people say, hear, see, and think.

• Plaud is a bootstrapped, skyrocketing, profitable company with a $250M revenue run rate achieved in just three years.
• Define the next-gen paradigm for human-AI interaction.
• Gain exposure to cutting-edge AI for Pro tools and play a direct role in our global expansion.
• Work with passionate teammates who value innovation, collaboration, and customer success.
• Grow your career in a culture that champions continuous learning and fast career development.
• Market-competitive compensation, global exposure, and a vibrant, creativity-fueled work atmosphere.

What you will do

• Own large-scale speech/audio data pipelines: build and maintain systems for data collection, cleaning, filtering, labeling, augmentation and quality control to support model training. Partner with senior speech engineers on term/hotword mining strategies based on ASR system output.

• Support model training and optimization: contribute to fine-tuning and evaluating speech/language models (SpeechLLM or general LLM background both applicable), with focus on improving recognition accuracy for code-switching, names, and product terms - working alongside domain specialists on the technical algorithm design.

• Domain adaptation: help fine-tune models using scenario-specific data to improve recognition of industry-specific terms across key languages and verticals.

• Evaluation: build test sets and evaluation frameworks (keyword/domain-lexicon based); benchmark internal models against open-source and commercial baselines.

Skills, qualifications and experience we look for

• Minimum 1 year of hands-on experience in speech, machine learning, or large-scale data engineering.

• Experience in AT LEAST ONE of the following:
  a) ASR/SpeechLLM model training, fine-tuning, or evaluation;
  b) LLM or general ML model training/fine-tuning;
  c) Large-scale audio, video, or text data pipeline work (e.g., tens of thousands of hours of audio, or TB-scale multimodal data).

• Solid Python and PyTorch fundamentals.

• Experience with distributed data processing (e.g., Spark, Ray) - moved up from nice-to-have, since responsibility #1 requires pipeline ownership at scale.

Nice-to-Have:

• Familiarity with SpeechLLM / speech SSL concepts, or exposure to models such as StepAudio or Qwen3-Omni.

• Prior experience building or contributing to hotword/contextual-biasing or code-switching ASR improvements.

• Publications at Interspeech, ICASSP, or other top AI venues, or speech-related patents.

• Experience owning a data workstream (tens/hundreds of thousands of hours of speech) in a speech-model training project.