Soumyajyoti Mahalanobish

Led the implementation of multi-tenant SaaS infrastructure using Terraform and ArgoCD, enabling automatic provisioning of on-demand customer environments. Ideated and implemented KEDA-based autoscaling for GPU fleets, optimizing resource utilization and reducing infrastructure costs by 53%. Implemented custom OpenTelemetry processors, enabling powerful correlation between metrics and traces.

Shipped core features for the Golang-based developer platform, including a user queue that reduced to and fro by 70% and improved testing efficiency for over 100 engineers. Mitigated security debts in software pipelines and initiated the move to Kubernetes external secret operator with secret rotations using AWS Lambda. Spearheaded the transition to self-hosted Github runners.

Implemented reliability for microservices on Kubernetes clusters by implementing Probes, Disruption Budgets, graceful termination, and service monitoring with kube-prometheus operator. Fine-tuned services, redefined HPA policies, integrated spot VMs, and created crons to scale down GKE node pools, resulting in a 41% month-on-month budget savings. Built robust DevSecOps pipelines with SAST, OS libr

Contributed to making microservices spin up efficiently by creating multi-stage dockerfiles, which decreased artifact size by 60%, and fine-tuning Kubernetes resource requests and limits. Scripted Ansible playbooks for a fleet of EC2 instances and integrated those with Hashicorp Packer templates.

Grade: 7.47

Completed a Bachelor of Technology degree with a major in Computer Science. Studied core concepts and principles in the field.