cn|
WorkWritingAboutConnect

cn

building things that matter

  • Work
  • Writing
  • About
  • Connect

© 2026 Cristian Najera

  1. Home
  2. /Work
  3. /AI-Powered Analytics Dashboard

Full-Stack Platform

AI-Powered Analytics Dashboard Real-Time Intelligence

Transforming raw telemetry into actionable intelligence for a Fortune 500 logistics company.

Client

Acme Logistics

Role

Lead Engineer

Year

2025

Duration

6 months

Project previewComing soon
Data Points

2.4M/day

Latency

<200ms

Uptime

99.97%

The Challenge

The client needed a unified view of fleet telemetry scattered across legacy systems, third-party APIs, and on-prem databases — with sub-second refresh for live ops dashboards.

The Approach

We built an event-driven pipeline with Kafka ingesting raw events, FastAPI serving transformed metrics, and a React dashboard rendering real-time charts via WebSockets.

Key Decisions

DynamoDB single-table design kept read latency under 10 ms at scale. Lambda@Edge handled auth without adding hops to the critical path.

Dashboard Load Time
12s→1.8s
Data Freshness
15 min→Real-time
Operational Cost
$48k/mo→$12k/mo
ReactFastAPIAWS LambdaDynamoDB
  1. 2024-07

    Discovery & architecture design

  2. 2024-08

    Pipeline MVP — Kafka + FastAPI

  3. 2024-10

    Dashboard v1 shipped to staging

  4. 2024-12

    Production launch with full fleet

The dashboard became our single source of truth overnight. We retired three legacy tools within the first month.

— Jordan Lee, VP of Operations, Acme Logistics
All projectsGet in touch →