Speech / Applied ML Engineer (Intern)

Role overview Improve speech recognition and related machine learning systems for diverse Southeast Asian languages, accents, code-switching and noisy audio. This internship spans experimentation, evaluation and production delivery, with attention to accuracy, robustness, latency, memory use and cost.

Responsibilities - Experiment with and tune speech recognition models for varied languages and accents. - Investigate errors and edge cases in imperfect, multilingual audio data. - Design evaluations that reflect production constraints and track model quality over time. - Improve inference efficiency and GPU use while considering latency and memory. - Work with engineering to integrate model changes into production pipelines. - Record experimental methods, results and tradeoffs.

Requirements - Experience with machine learning model training, fine-tuning and evaluation. - Familiarity with GPU inference and optimization workflows. - A practical engineering approach that considers how models behave in production. - Initiative to investigate failures, communicate uncertainty and follow through on debugging. - Ability to balance careful experimentation with delivering useful improvements.

Nice to have - Experience with multilingual, low-resource or regional speech data. - Exposure to on-device or low-latency inference. - Experience deploying machine learning models in production.

Benefits and work setup - Hands-on experience with applied machine learning under real production constraints. - Collaboration with senior technical colleagues and the opportunity to build a portfolio of experiments and deployed improvements.