Case study 02 · Sanitized private system
Daily Momentum Command Center
A private task platform that keeps browser actions, Telegram commands, reminders, and scheduled check-ins on one durable domain model.
Product workflow, API, persistence, messaging interface, reminders, deployment, and tests
The problem
What needed to change
Tasks entered in chat, reminders, and a browser dashboard easily drift into separate sources of truth. The system needed one durable task model, consistent behavior across interfaces, and reliable reminder semantics.
I designed and built the API, SQLite task store, typed domain models, Telegram command layer, reminder workflow, dashboard actions, private deployment, and automated tests.
System map
One request, traced end to end.
- 01Interfacesbrowser UI · Telegram
- 02FastAPItyped requests · domain actions
- 03TaskStoreSQLite lifecycle + reminders
- 04Schedulercheck-ins · due reminders
- 05Private accesstoken gate · tailnet bind
Key decisions
Designed for trust and operability.
Make every interface use the same state
The browser UI, API, Telegram commands, check-ins, and reminder sender all use one SQLite-backed TaskStore instead of maintaining interface-specific copies.
Model lifecycle behavior explicitly
Create, update, complete, snooze, and permanent delete have distinct semantics. Reminder delivery records the send and advances the schedule instead of silently duplicating notifications.
Constrain the API boundary
Typed Pydantic models validate areas, priorities, statuses, time fields, completion criteria, materials, location context, and estimates before data reaches storage.
Keep the production footprint appropriate
A supervised user service binds to a private network interface, uses a token gate, persists locally, and exposes a health endpoint for operational verification.
Tradeoffs
What this design optimizes—and what it does not.
- SQLite is appropriate for a single-user service but not a multi-writer SaaS deployment.
- A lightweight command parser favors predictable behavior over unrestricted natural language.
- Private-network deployment reduces exposure but intentionally limits public demos.
Verification
Private does not mean unverifiable.
- 38 automated tests cover the private system.
- CRUD, snooze, reminders, Telegram parsing, dashboard, API, and scheduler behavior are exercised.
- Health checks and direct database readbacks support release verification.
- The same state is exercised through browser and messaging workflows.
Interview walkthrough
Ask me to trace the system live.
I can provide a guided, sanitized walkthrough of the workflow, representative code, architecture, automated tests, failure modes, and the changes required at larger scale.