← Index
← Back to selected work
Case 03 / Grab · 2024

From hours-delayed
to live.

Grab had IoT event streams generating data across Southeast Asian markets but no unified way to see operational performance in real time. Leadership was making decisions on data that was hours stale. The process existed — the visibility didn't.

RoleArchitect · pipeline lead
RegionSoutheast Asia · multi-market
StackAzure DF · Synapse · Power BI
OutcomeLive executive dashboards
01 · The constraint

Decisions in a market like SE Asia move fast. When operational data arrives hours after the fact, you're not running the business — you're describing it. Grab's leadership needed to see what was happening, not what had happened.

02 · The pipeline

Azure Data Factory for ingestion, Azure Synapse Analytics for transformation, Power BI for live executive dashboards. I led architecture-review sessions with Grab's engineering team across markets and validated throughput & latency requirements before go-live.

01 · ingest
IoT event streams
Multi-market raw events landing continuously.
02 · transform
Synapse
Schema, joins, market-aware enrichment.
03 · serve
Curated layer
Operational truth surface, query-ready.
04 · see
Power BI live
Executive dashboards across SE Asia markets.
03 · What changed

Real-time operational visibility across Southeast Asian markets. Decisions shifted from hours-delayed data to live operational insight. Documentation went to the customer team detailed enough that the pipeline survived after my engagement ended.

Live
Executive dashboards across markets
Hours → seconds
Operational latency on key KPIs
Multi-region
Architecture reviewed with Grab eng team
Self-run
Customer team independent post-engagement
“The process existed. The visibility didn't.
I built the visibility.”— how I framed it
Back to first case
Walgreens — AI supply chain across 200+ stores