About
The ML startup faced high costs during its growth for a data-driven platform infrastructure that processes around 30 TB per month and stores raw data for 12 months on AWS. We reduced the monthly cost from $75,000 to $22,000 and achieved 30% performance over SLA.
Challenge, approach, and impact
High-load Infrastructure Scalability
The client needed an AWS-based infrastructure capable of handling over 2,000 queries per second while maintaining high performance and stability. The existing setup could not reliably support the required traffic growth, creating risks of performance degradation under peak loads.
High Infrastructure Costs
Monthly infrastructure costs reached approximately $75,000, making the solution financially inefficient. The client required significant cost optimization without compromising system performance, availability, or scalability.
Reliability and Fault Tolerance
The system required 99.9% availability and strong fault tolerance. The existing infrastructure lacked sufficient replication and resilience mechanisms, increasing the risk of downtime and service interruptions during failures or high-load scenarios.
How we built
Testimonials
Robert P.
Deduce
“They have very intelligent people on their team — people that I would gladly hire and pay for myself.“
Team structure
Client team
Robert
CTO
Project stakeholder
The client stakeholders at Deduce were working closely with the team at Dataforest
Agency team
2 x Product Manager
Production
1 x Business Analytics
Production
2 x Data Engineer
Production
