Where recommendations are used
NativePath may recommend a specific lesson or next learning step based on the learner's saved progress and assessment history.
For example, repeated mistakes in a task or quiz may trigger a suggestion to review an earlier published lesson related to that concept.
Information used
The system may use the selected learning route, course/lesson/task/quiz identifiers, assessment outcome, repeated-error count, score/status, event time, inferred learning concept, and later mastery evidence.
These signals come from the learner's activity inside NativePath and server-side assessment results. Raw learner code is not stored inside the recommendation learning-signal record.
How a recommendation is selected
A single mistake should not create a persistent remediation recommendation. The current logic requires at least two relevant signals for the same concept within its analysis window.
NativePath then looks for a relevant previously published lesson. Verified later mastery resolves the old signal history; new repeated mistakes are required before the recommendation may appear again.
AI and authority
AI may explain or classify an already saved failure, but it does not determine authoritative PASS/FAIL, manual-review outcomes, or certificate eligibility.
Recommendations are advisory learning guidance and do not themselves change assessment or certification status.
Contact
Resource owner: ИП Кожин Андрей Сергеевич. Legal messages about recommendation technology may be sent to hello@nativepath.ru.