The stack behind structured writing practice
Jent Ulvak runs on a deliberately chosen set of tools. Each component was selected because it solves a specific problem in delivering writing instruction at scale — not because it was popular at the time.
- 1 Adaptive video delivery Lessons stream through an adaptive bitrate layer that adjusts to connection quality. Students in regions with variable bandwidth receive the same content without buffering interruptions.
- 2 Asynchronous feedback engine Written exercises are submitted through a structured form system. Instructors annotate directly on submitted text using a comment-threading model rather than external email chains.
- 3 Progress tracking layer Each module completion is recorded against a learner profile. The system surfaces incomplete exercises and pending feedback without requiring students to navigate manually.
How the tools compare across delivery needs
The table below maps each platform component against the specific requirements of writing instruction. Gaps are marked plainly — no component does everything.
| Component | Async learning | Live feedback | Offline access | Annotation |
|---|---|---|---|---|
| Video module system | ||||
| Submission portal | ||||
| Progress tracker | ||||
| CDN delivery layer | ||||
| Encrypted storage |
Why these choices matter
Writing instruction depends on the quality of the feedback loop, not just content delivery. The submission and annotation tools were prioritised over live session infrastructure because most writing errors surface in the revision process, not during initial instruction. A student reviewing annotated drafts at their own pace retains more than one attending a live correction session without preparation. The current stack reflects that ordering of priorities — and will be updated when evidence suggests otherwise.
Security by default
All student submissions are encrypted at rest and in transit. Instructor access is role-scoped — no administrator can read student work without an explicit course assignment. Jent Ulvak has operated under this policy since 2022.
Latency across regions
Video content is distributed across edge nodes covering Ukraine and neighbouring countries. Average load time for a 720p lesson segment is under 1.8 seconds on a standard 4G connection — measured quarterly, not estimated.
Maintenance and updates
Platform components are updated on a fixed cycle. Students receive advance notice before any interface change. Downtime for maintenance is scheduled outside peak learning hours — typically between 02:00 and 04:00 Kyiv time.