AI Spoken Language Assessment Developer

Klavio Works

Remote · $80k – $100k / year · Full-time · Posted 4 months ago

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

About the role

Automated speaking assessment is one of the hardest problems in language learning technology. A learner records a sentence. You need to evaluate pronunciation accuracy, fluency, grammatical correctness, and communicative effectiveness — all from a noisy audio recording, often from a non-native speaker in an accent the system was not trained on. We build the assessment engine for a language learning platform with 4 million active learners across 18 native language backgrounds learning English, Spanish, and Mandarin. The current pronunciation feedback system scores overall quality but cannot tell learners which specific sounds they are mispronouncing. That granular feedback is what drives actual improvement. You will build the phoneme-level pronunciation assessment: the forced alignment pipeline that maps learner speech to target phoneme sequences, the scoring models that evaluate pronunciation quality per phoneme, and the feedback generation that tells learners specifically what to practice.

Fully remote.

Responsibilities

Build forced alignment pipeline mapping learner speech to target phoneme sequences
Develop per-phoneme pronunciation quality scoring models
Create feedback generation system explaining specific pronunciation errors
Handle cross-accent variation for learners from 18 native language backgrounds
Validate assessment accuracy against professional teacher evaluation ground truth
Build evaluation metrics correlated with learner improvement outcomes

Requirements

3+ years in speech ML with pronunciation assessment or spoken language evaluation experience
Experience with forced alignment tools: Montreal Forced Aligner, Kaldi, or similar
Strong PyTorch skills and familiarity with speech processing libraries
Understanding of phonetics and cross-lingual pronunciation challenges
Ability to collaborate with linguists and pedagogical experts