The problem
AccioJob's infrastructure had grown organically on GCP: managed services billing per request, over-provisioned databases, a separate DevOps function, and AI workloads whose inference costs scaled linearly with usage. Projected annualized spend at enterprise scale was north of $200K.
The re-architecture
Neeraj ran a complete multi-cloud migration from GCP to Azure and AWS, solo, with no external DevOps support. The moves that mattered: placing workloads where reserved pricing and credits were best, replacing per-request managed services with self-hosted equivalents, aggressive storage tiering, LLM model-routing to cut inference spend, and cross-training the team so a separate DevOps function was no longer needed.
In parallel, he secured $300,000 in cloud credits, $200K via Microsoft for Startups and $100K via AWS Activate, by showcasing high-volume production AI usage.
The result
Enterprise-scale infrastructure now runs at a fraction of its former cost, a reduction of more than 70%. The same playbook (audit, re-architect, pursue credits) is what Axionry sells as Cloud Cost Optimization.