B2B

Python

remote

Assignment description:

Our Customer is building next-generation products for math learning, putting AI at the core of their solution. Operating as an early-stage, flat team of 6–7 people, they prioritize high autonomy, fast execution, and zero bureaucracy.

You will work on a greenfield project building an adaptive math learning application targeted at 7–12-year-olds for B2C (with potential B2B school expansion). The product uses Socratic teaching techniques, speech-to-text/text-to-speech interactions, and general LLM models to guide students through customized study flows. Rather than building models from scratch, you will design the surrounding product, APIs, and data architecture.

A key part of the work involves learner signal engineering—analyzing real-time user signals to tune adaptive logic so the system responds dynamically to how fast a child learns. AI coding agents are a core part of the daily workflow; you will direct, shape, and critically review agent-generated code to ship fast, high-quality features.

About the team:

A senior, tight-knit group of 6–7 engineers operating without middle management layers. AI coding agents are embedded into the primary development process rather than treated as a side experiment. You will report directly to the COO/CTO, enjoying direct ownership over your domain.

Must haves:

- Senior-level proficiency in TypeScript across backend, web, and native/mobile environments.

- Genuine, practical experience using AI-agentic tools as a primary daily coding methodology (hard requirement).

- Solid hands-on Python experience for data handling and signal analysis.

- Proven track record building robust backend APIs, database architecture, and native application features.

Other requirements:

- Strong interest or background in learner signal engineering, adaptive algorithms, or parameter tuning for personalization.

- High self-direction and ability to thrive in a fast-moving, greenfield environment.

- Strong critical-review skills for evaluating agent-generated code for security, performance, and correctness.

About the customer:

Our Customer is building next-generation products for maths learning across both home and school environments. They are developing a greenfield, adaptive learning platform for kids aged 7–12. Built alongside their flagship product, this from-scratch initiative leverages data and AI to personalize study flows, adjust learning speeds, and deliver tailored educational challenges.