December 1, 2024
Everest Railcar Tracking Dashboard
Production freight rail SaaS with real-time visibility, a 3-queue BullMQ pipeline, automated exception engine, SPLC resolution, and 9 live KPI widgets.
- Node.js
- TypeScript
- PostgreSQL
- Redis
- BullMQ
- Prisma
- Trimble Maps
Problem
Freight rail operators depended on manual daily CSV trace report processing. Exception review was slow, location resolution across railroad city codes was incomplete, and fleet managers lacked live KPI visibility on an interactive map.
Architecture decisions
- Designed a 3-queue BullMQ pipeline (
csv-import,exception-evaluation,od-recalculation) so large CSV ingest never blocked the main API thread. - Used idempotent job IDs and atomic Prisma transactions for exactly-once ingestion.
- Modeled 5 exception rules with per-customer threshold overrides and
ExceptionClearanceaudit trails. - Integrated 214,000+ SPLC codes for precise railroad city-code resolution.
- Secured the API with JWT + Passport, Zod validation, OpenAPI docs, and Pino/Morgan observability.
100%
Manual CSV processing eliminated
Daily trace reports fully automated
214k+
SPLC location codes resolved
Railroad city code database
5
Automated exception rules
Bad Order, Dwell, Transit Delay, Underutilization, Bad Customer
9
Live KPI widgets
Fleet status, exceptions, utilisation, ETA
Role and impact
Owned backend architecture and delivery for the production SaaS used by freight rail operators. Automated daily CSV ingestion end to end, caught railcar issues earlier than manual review, and gave fleet managers live dashboard visibility they did not previously have.
Tech stack
- Backend: Node.js · TypeScript · Express.js · BullMQ · Prisma ORM
- Database: PostgreSQL · Redis
- Mapping: Trimble Maps API
- Infrastructure: AWS (S3, EC2) · Docker · GitHub Actions CI/CD · Nginx
Related projects
5 Peaks Youth Solutions
Replaced spreadsheet program tracking with FERPA-aware portals, dual KPI engines, and a minimum-cohort privacy guard.
Azara Health Care Portal
Delivered 4-role RBAC dashboards and intelligent matching that cut caregiver assignment from manual search to ranked matches in seconds.