Client
Magnum
Selected work / Magnum
A notification layer that keeps employees on top of training and gives HR a way to talk to the right group without spamming the whole company.

Client
Magnum
Channels
Telegram · WhatsApp
Sources
Moko · Magnum Space
Magnum
Two systems running on the same platform: training notifications driven by the LMS, and a segmented broadcast tool for internal communications.
The LMS does not push events. So the system had to detect change itself: regular sync, status diff, then the right reminder to the right person at the right time.
LMS sync, training reminders with manager escalation, an admin panel for targeted broadcasts by branch, department or office, and delivery over Telegram and WhatsApp.
Case study
Two notification streams on one backend. The first one is tied to the LMS — Moko and Magnum Space — so employees actually finish the training they have been assigned. The second one is a broadcast tool HR uses to talk to a specific branch, department or office instead of the whole company. What makes the project interesting is the LMS side. The platforms do not emit events, so all of the "you have a new course" / "your deadline is in three days" logic has to be reconstructed by the notification system itself.
Problem
The LMS could assign courses but it could not chase anyone. Employees would forget. Managers had no easy view of who was behind. Deadlines slipped, often quietly. On top of that, the company had no clean way to send a message to "everyone at this branch" or "everyone in this department". HR had to coordinate that by hand across managers and chats.
Solution
We sync the LMS on a schedule, diff the state, and generate the events the platform never emitted. From there a reminder rules engine takes over: a message when a course is assigned, follow-ups every seven days while it is still open, a completion notice when it is done, and an escalation to the employee's manager when a deadline passes. On the broadcast side, an admin panel lets HR pick the audience — a branch, a department, an office, a phone list — and ship a message to Telegram and WhatsApp at once.
Technical work
The LMS gave us status snapshots, not events. So the heavy lifting is on our side: pull frequently enough to be useful, diff cheaply, and treat the diff as the real source of truth for everything downstream. Once the events are clean, the reminder rules and the channel routing are straightforward. The broadcast side reuses the same delivery layer. The admin panel turns "this department, this office" into a phone list that flows into the same Telegram + WhatsApp pipeline.
An AI HR agent for a Kazakhstani retail chain: candidate screening, internal knowledge, vacancy fit, and recruiter workflows.
An internal AI assistant for Magnum employees: ask a question in plain language, get an answer pulled from the company knowledge base.
An AI support agent for one of the largest running events in Central Asia, answering participant questions and routing edge cases.
We help operations and HR teams put a real notification platform behind their daily work: LMS sync where the LMS will not push, segmented broadcasts, and clean delivery to the channel employees already use.