Principal Machine Learning Engineer at Speechmatics

As a Principal Machine Learning Engineer, you will co-define Speechmatics’ technical vision while remaining hands-on in the Modelling Team. You will bridge the gap between cutting-edge research and production systems, developing next-generation transformer models and optimizing inference for global scale. This is a high-ownership role focused on shipping real-world impact.

• Develop and deploy advanced ML models using Python and PyTorch, translating research into scalable, maintainable production services and services.
• Optimise large-scale model inference using strategies like dynamic batching, flash attention, and speculative decoding to improve speed and cost efficiency.
• Define and enforce best practices for model lifecycle management, data quality, and evaluations across the entire ML stack.

• Deep expertise in building and shipping production-grade ML systems, with a strong foundation in modern transformer architectures and self-supervised learning.
• Proven track record in distributed training and optimizing inference at scale, bridging the gap between research models and production-ready code.
• Expert proficiency in Python and ML frameworks such as PyTorch, complemented by experience in MLOps, CI/CD pipelines, and containerization.

• Lead the technical direction of a $62M Series B scale-up that recently saw 4x growth in real-time usage and serves global customers.
• Work at the rare intersection of high-rigor ML research and large-scale production, shipping models that solve real-world problems like accents and noise.
• Take full ownership of technical domains, raising the bar for a talented team of 10+ engineers while developing groundbreaking bilingual and medical models.