Human-led lessons
Teacher-recorded explanations and question walkthroughs, organised into focused subtopics.
Make sense of STEM, one topic at a time. Learn with a teacher, follow the working, and practise at your own pace.
Chemistry prototype available · Built by a teacher and students
Reacting masses · Original worked example
Our starting hypothesis: students need a clearer path from understanding a concept to applying it independently in an exam-style question.
Lessons, worked solutions and practice often live in different places. Students have to build their own learning journey.
Students need to see how an equation, a calculation and a scientific explanation fit together.
Teachers need a simple view of what students completed and how they performed between lessons.
The MVP will connect these foundations. This submission demonstrates one Chemistry topic with sample media and local progress tracking.
Teacher-recorded explanations and question walkthroughs, organised into focused subtopics.
Step-by-step explanations using scientific terminology, units and relevant methods. Teacher review planned before student release.
Original exam-style questions by topic. Separate topical and yearly paper access subject to permissions.
Student accounts and saved progress in the MVP. The prototype uses a clearly identified local demo profile.
Lesson completion, practice attempts and scores in a straightforward dashboard.
Resources mapped to subject, level and syllabus version as coverage expands from Chemistry to STEM.
One educator and three student team members. The roles below define our responsibilities for this project; student roles are development opportunities, not claims of prior professional experience.
9+ years teaching Cambridge IGCSE and AS/A Level Chemistry. Leads curriculum decisions, lesson design, scientific review and the pitch.
Academic quality plan ↗Organises topics and example questions, checks explanations from a learner's perspective, and supports captions and lesson preparation.
Learning experience ↗Collects student feedback, tests the learning journey, documents issues and coordinates the demo story and presentation rehearsal.
Demo presentation ↗Owns the technical checklist, interface testing, deployment checks and future API integration, with AI-assisted development support.
Technical plan ↗Founder’s Cambridge Chemistry teaching experience.
Our founder brings subject expertise; our student team brings direct insight into how learners use explanations and practice. Founder-reported access to marked paper scans across 14 presidential schools is a potential research resource, subject to authorisation, de-identification and review. These schools are not presented as partners or customers.
Cambridge STEM scope, teacher-led learning and a focused Chemistry demonstration.
Working lesson-to-practice flow, sample explanations, local progress and a teacher view.
Real accounts, database, API-supported explanations and an initial teacher-reviewed content collection.
Paid offering, reliable support and staged subject expansion, guided by learning outcomes and retention.
*Indicative planning window, dependent on team capacity, content production and funding. Not a committed launch date.
Use an external AI API for the MVP. Develop our own hosted, customised open-weight model only after the data, evaluation and operating costs justify it.
| Stage & owner | What we build | Technology & quality gate |
|---|---|---|
| Prototype Dustbek + team | Lesson playback, practice scoring and linked demo progress. | HTML, CSS, JavaScript, browser storage, SVG and MP4. AI-assisted development; no live AI endpoint in this version. |
| MVP Dustbek + Keyhan | Accounts, persistent records, teacher access and AI explanations grounded in approved content. | Planned: web backend, relational database, managed authentication and server-side AI API. Teacher-reviewed content before release. |
| Content & pilot Keyhan + Rasulbek + Shohruh | Reviewed lessons and questions, usability sessions and comparable pre/post assessments. | Content versioning, feedback logs and an evaluation set. Measure retention, correctness and teacher time saved. |
| Independent AI Future technical capacity | Extract and curate authorised scans; evaluate and adapt an open-weight model. | OCR/document extraction, human checking, separate training/evaluation sets, model licensing review and hosted inference. |
Beyond the MVP: detailed lost-mark feedback, a hint-based tutor, a mistake notebook with spaced revision, and teacher assignments. These are future features, not active prototype capabilities.
Test student subscriptions and school licensing through paid pilots. Pricing and financial forecasts will follow customer research.
Track independently assessed improvement, continued use, teacher time saved and delivery cost per active learner. We are seeking pilot feedback, technical mentorship and investment conversations.
Try the Chemistry prototype or watch our short overview.