3 months ago
Base Salary
$155k - $205k/yr
Responsibilities
- Define the impact of AI on children’s education.
- Own AI systems and features end to end, from identifying product opportunities through research, evaluation, deployment, and production improvements.
- Build infrastructure to measure educational quality, engagement, safety, latency, and AI-teacher behavior.
- Design and improve the tutoring agent harness, including teaching decisions, intervention timing, and adaptation to each child.
- Develop learner profiling and adaptation systems that model children’s strengths, gaps, pace, and engagement patterns.
- Create feedback loops and data flywheels that improve product quality through usage, evaluation, and experimentation.
- Collaborate with learning experts and product engineers while shipping AI features into a real product.
Requirements
- At least 3 years of experience building AI products in an engineering or research role.
- Strong software engineering fundamentals and proficiency writing clean Python code.
- Experience building and operating evaluation or measurement systems, including AI evaluations.
- Ability to translate ambiguous product-quality questions into metrics, research plans, experiments, and production improvements.
- Understanding of machine learning foundations, including the ability to read recent ML papers and implement parts of the ML stack.
- Applied mathematics and linear algebra proficiency sufficient to contribute to research discussions.
- Experience leveraging AI in engineering workflows and curiosity about how AI is transforming work.
- Public technical projects, strong product intuition, low-latency inference at scale, model training experience, and familiarity with speech systems are bonus qualifications.
Benefits
- In-person work from the San Francisco office.
- Potential US visa sponsorship depending on the candidate’s situation, including O-1 visa support.
- Collaborative work with a 5-person applied ML team and advisors from Stanford and leading industry research labs.
Tech Stack
Jupyter NotebookPython
Categories
About Ello
Building a 1:1 teacher was one of the original dreams of AI. It’s finally possible.