Principal Machine Learning Engineer at Speechmatics

You will lead the translation of cutting-edge research into scalable, customer-centric speech solutions. This high-ownership role involves developing diverse ML models, including large self-supervised architectures, to maintain the world’s most accurate speech recognition system. You will span the entire codebase, driving technical vision and mentoring engineering teams to achieve excellence in model deployment.

• Develop and optimize advanced ML models for speed, accuracy, and cost efficiency, ensuring seamless transition from research to production.
• Define and enforce technical best practices across the organization for code quality, model lifecycle management, and containerized deployment.
• Mentor fellow engineers and foster a culture of technical excellence while identifying and solving complex ML bottlenecks across the entire product stack.

• Possesses deep expertise in the modern ML stack, including contemporary transformer architectures (GQA, KV-caching) and distributed training techniques.
• Demonstrates a proven track record of deploying large-scale models using Python, PyTorch/TensorFlow, and MLOps tools like Docker and Kubernetes.
• Brings a specialized background in speech-to-text or text-to-speech technology, ideally with contributions to open-source projects or top-tier research publications.

===

• Work at the forefront of Voice AI innovation, building models that help the world "Understand Every Voice" through world-leading speech technology.
• Take on a high-impact technical leadership role with total ownership, co-defining the long-term technical vision for an organization at the peak of its market.
• Join a specialized modeling team that balances high-level research with robust engineering practices, utilizing the latest in transformer architectures and MLOps.