Machine Learning Engineer - Singapore
Core
Own large-scale speech and audio data pipelines for collection, cleaning, filtering, labeling, augmentation, and quality control; support fine-tuning and evaluation of speech and language models to improve recognition of code-switching, names, and product terms.
Role type
Machine Learning Engineer (Speech/Audio)
Builds
Large-scale speech and audio data pipelines, evaluation frameworks, and fine-tuned speech models
Domain
Speech technology, Audio processing, Natural Language Processing
Deliverable
production ML models
Required skills
Python, PyTorch, Distributed data processing (Spark, Ray), ASR training and evaluation, Large-scale data pipeline engineering, Model fine-tuning, Data cleaning and labeling
Preferred skills
SpeechLLM concepts, Self-supervised learning (StepAudio, Qwen3-Omi), Hotword mining, Code-switching ASR improvements, Speech-model training data ownership, Publications or patents in speech technology
Technologies
PyTorch, Spark, Ray, StepAudio, Qwen3-Omni
Responsibilities
Build keyword- and domain-lexicon-based test sets and evaluation frameworks; Benchmark internal models against open-source and commercial baselines; Partner with senior speech engineers on term and hotword mining strategies
Seniority
Junior to Mid-level, hands-on IC
