Build innovative, high-quality language assessments using cutting-edge NLP.
Duolingo AI Research is a nimble and fast-growing team that has already revolutionized language learning for more than 300 million people. We are looking for a talented researcher/engineer with creative interdisciplinary ideas to help us deliver next-generation language assessments to anyone, anywhere, through AI.
As an ML/NLP Engineer, you will apply your background in text and speech processing to create state-of-the-art automated scoring methods for language assessment. These systems will push the boundary of what personalized, adaptive learning and proficiency testing can be. You will work with a diverse team, come up with novel hypotheses, test them, and then scale them up to production systems unlike what you’ll see anywhere else. You will work with the award-winning Duolingo English Test (DET) team, and support our efforts to advance Duolingo’s mission of educational access through science.
Apply ML and NLP methods to large data sets
Prototype new models for analyzing text and speech from both native and non-native speakers, evaluate with small scale experiments, and productionize solutions at scale
Work with a cross-functional team of award-winning engineers, linguists, and others to build new product features
Iterate on product quality through continuous A/B testing
Graduate-level expertise (or equivalent industry experience) in machine learning, natural language processing, signal processing, or related field
Experience implementing ML systems at scale in Python, Java, Scala, or C/C++ (i.e., not just R or MATLAB)
A strong mathematical background in statistics and machine learning
Great presentation and communication skills
Ability to relocate to Pittsburgh, PA
EXCEPTIONAL CANDIDATES WILL HAVE
MS or PhD in machine learning, natural language processing, or related field
Previous experience building NLP systems for commercial products or government projects, such as speech recognition or machine translation (automated scoring for language assessment a plus!)
A portfolio of relevant publications or open-source projects to share with us
A desire to keep up with the field by attending or publishing at relevant conferences (ICML, NeurIPS, ACL, EMNLP, ICLR, Interspeech, etc.)