Selected Work
2025 · Cloud
Verdant Climate
A sensor ingestion pipeline processing 4M+ IoT readings/hour across 11 regions. Cost-optimized to 38% of original cloud bill.
Verdant Climate
Client
Verdant
Role
Cloud Architect
Year
2025
Discipline
Cloud
(01)
Problem
Verdant's climate sensors were generating so much data that their AWS bill hit $84k/month — and they were still losing 2% of readings during peak hours. They needed to scale 10x without scaling cost.
(02)
Solution
Redesigned the ingestion pipeline around Kafka + TimescaleDB with tiered storage (hot → warm → cold), replaced Lambda with Kinesis + containerized workers, and built a Terraform module that auto-scales per region. All observable via Grafana.
(03)
Outcome
Monthly cloud bill cut from $84k to $32k (62% reduction). Zero data loss in 8 months. Pipeline now handles 10x the original load with headroom.
-62%
Cloud cost cut
4M/hr
Readings processed
0
Data loss events
11
Regions deployed
Stack
AWSKafkaTerraformTimescaleDBGrafana
Next case