About the job
Join Our Team as a Machine Learning Engineer for Assessments
At Speak, we're on a mission to transform the language learning experience.
Learning a new language has the power to enrich lives, facilitating connections with diverse cultures, career opportunities, and communities. With two billion people worldwide engaged in language studies, the traditional one-on-one tutoring approach remains challenging to scale and has seen little innovation over the years. Speak is revolutionizing this space by offering an AI-powered, human-like tutoring experience that prioritizes conversation. Our platform allows learners to practice speaking, receive instant feedback, and progress through expertly crafted lessons, ensuring a seamless journey from beginner to confident speaker in multiple languages.
Since our inception in South Korea in 2019, Speak has rapidly ascended to become the leading language learning app, serving learners across various markets with 15+ languages. Backed by over $150 million in venture funding from prominent investors such as OpenAI, Accel, and Khosla Ventures, our distributed team spans San Francisco, Seoul, Tokyo, Taipei, and Ljubljana.
About the Role
We are seeking a talented Machine Learning Engineer for Assessments to spearhead the development of top-tier assessment systems across our diverse product lines, including Speak for Business and B2C offerings. You will collaborate closely with our Assessment Design Lead, along with teams in Machine Learning, Product, and Engineering, to translate assessment frameworks and rubrics into robust, scalable scoring and feedback systems.
This role will encompass the implementation, deployment, and continual enhancement of our assessment algorithms and ML systems. While immediate focus will be on refining and expanding current assessments, the work will also contribute to a foundational capability that can be leveraged across our platform.
Your Responsibilities
- Lead the development of assessment ML systems end-to-end
- Design, deploy, and maintain scoring models and pipelines (feature extraction, model training, inference, feedback generation).
- Oversee monitoring, regression tests, and iterative improvements to ensure accuracy standards are met.
- Establish and implement evaluation frameworks
- Create validation and evaluation structures for assessments, incorporating metrics, test sets, and both offline and online analyses.
- Convert assessment needs into quantifiable acceptance criteria and safeguards.