Assessment Design Lead

Speak

Remote regions

Global

Benefits

Similar Jobs

See all

Role Overview:

  • Own assessment design for three types: Curriculum Mastery, Proficiency, and Placement Tests.
  • Define constructs like fluency, pronunciation, grammar, and task achievement with clear rubrics.
  • Shift from a vague goal to a concrete, scoreable system with the ML engineer.

Responsibilities:

  • Define constructs and build rubric/blueprint layers for item writers and scoring models.
  • Own validity and quality bar, signing off on content validity and auditing for bias.
  • Run validity evidence plans with concurrent/criterion studies against external measures.

Collaboration:

  • Partner tightly with Product Manager and ML Engineer on automated scoring and calibration.
  • Keep construct and quality bar separate from model implementation.
  • Ensure tight loop between construct definition, rubric, and shipped scoring model.

Speak

Speak is building a human-level AI-powered tutor for language learning, providing a conversation-first experience that lets learners speak, get feedback, and progress through lessons. The company has raised over $150M from investors like OpenAI and Accel, with a distributed team across San Francisco, Seoul, Tokyo, Taipei, and Ljubljana.

Apply for This Position