Skip to content
Search jobs inside your AI agent. Claude, ChatGPT, Claude Code, Codex, Gemini, Cursor and Grok
Connect

Alpha

NewBe an early applicant

Speech-Language Pathologist - Project Contract

Posted

Description

A five-year-old who says "thun" for "sun" is not misreading. You can hear that in two seconds, and this project puts that ear where it matters, helping a reading app understand thousands of kids at once.

Alpha School is a private school network where students master core academics in about two hours a day with AI tutors. Its youngest students learn to read in AlphaPhonics, an app that listens to them sound out letters and words. The app needs an expert ear for how 4- to 6-year-olds really speak, including accents and developing speech.

The fix is data, and judgment the data doesn't have. A team of teachers is labeling about 40,000 recordings of children reading. You set the standard they label to. First, you write the guide and the reference examples, and you record a clear articulation video for every English phoneme. Then you spot-check the team's work and label the hard cases yourself in IPA. As the model retrains, your job shifts to reviewing what it still gets wrong and telling engineers why. A typical week: an hour fixing a guide rule the labelers keep splitting on, a batch of tricky /r/ and /l/ clips, and a call explaining to an engineer why a lateral lisp should not count as a misread.

This is not clinical work. There is no caseload, no IEPs, and no reports. It is fully remote, about 10 hours a week, on your own schedule. If you've ever wished the reading apps your students use understood how young children actually talk, apply and help build one that does.

What you will be doing

  • Creating articulation videos for all English phonemes to support an early reading app
  • Labeling student speech samples at the phoneme level (IPA), with careful attention to typical and atypical speech patterns in 4- to 6-year-olds
  • Distinguishing age-appropriate developmental errors from true misreads, so the app responds the way a skilled early-childhood SLP would
  • Writing and maintaining labeling guidelines so the dataset stays consistent across samples and annotators
  • Reviewing the app's pronunciation-scoring outputs, flagging misses, and explaining them in terms engineers can act on
  • Advising the product team on speech-sound norms and developmental progressions

Responsibilities

Make sure Alpha's reading app hears 4- to 6-year-olds the way a skilled pediatric clinician would.

Requirements

  • Master's degree in speech-language pathology and current ASHA CCC-SLP
  • Hands-on experience with articulation and phonological development in children ages 4 to 6
  • Comfort working in AI tools and web-based labeling software
  • Available about 10 hours a week for 3 to 6 months

Nice to have

  • Prior speech-data annotation or labeling for AI or speech-recognition systems
  • Experience with bilingual children or dialects such as African American English and Spanish-influenced English
  • Background in early literacy, phonics, or the science of reading
  • Experience writing clinical protocols, rubrics, or training materials for other clinicians or aides
Apply

See also

Healthcare jobs by country — openings, pay and top skills →