Subroto Kumar
Software Engineer with 2.5 years building and operating cloud-native systems at production scale on AKS and EKS. Specializes in Kubernetes, Distributed Systems, Cloud Architecture, DevOps, and Multi-Cloud engineering.
01. Architecture & Identity
I am a Software Engineer specializing in DevOps, Cloud & MLOps Engineering, Kubernetes Administration, and Distributed Systems. My expertise lies in designing scalable, event-driven microservices and shipping zero-downtime GitOps pipelines across multi-cloud environments like AKS and EKS.
Beyond traditional infrastructure, I architect robust backend systems that power modern AI, from highly optimized RAG platforms to provider-agnostic LLM gateways. I build resilient foundations that hold up under production load, ensuring every system is scalable, observable, and built right. 🚀
$ get-certs --vendor="Microsoft Azure"
✔ AI-102: AI Engineer Associate
✔ AZ-900: Azure Fundamentals
02. Execution Log
Senior Software Engineer
- Owned end-to-end GitOps delivery on AKS/EKS: authored Helm charts, configured ArgoCD sync policies, and enforced GitLab CI gates with Trivy and SonarQube, achieving 25% faster rollouts with full uptime across 12 months.
- Eliminated config drift across 100% of production clusters by migrating NGINX Ingress to NGINX Gateway Fabric via auditable IaC, standardizing TLS termination and cutting routing incidents by 80%.
- Built a provider-agnostic API gateway in Go unifying Azure OpenAI, AWS Bedrock, and Ollama behind one REST interface, enabling zero-code model switching and saving 40+ engineering hours per integration cycle.
- Designed a Kafka-backed event-driven Spring Boot microservice routing 3,000+ agent incident tickets from ServiceNow via consumer-group partitioning, sustaining full message durability and cutting manual dispatch by 60%.
- Deployed Kubeflow Pipelines on AKS for ML workflow orchestration, standardizing model training DAGs across 3 data science teams and cutting experiment turnaround by 35%.
- Scaled a RAG platform from 0 to 17,000+ MAU on AKS by tuning HNSW parameters on PGVector and building a custom query planner, sustaining sub-50 ms p99 latency under production load.
- Instrumented a platform-wide observability stack on OpenTelemetry and Grafana tracking p99 latency and SLO burn rates, compressing MTTD from hours to under 5 minutes.
- Rolled out a fault-tolerant Azure Functions ingestion pipeline for 5+ file formats with exponential-backoff retries and dead-letter queue recovery, maintaining complete data integrity across all production runs.
- Mentored 2 junior engineers on Kubernetes operators, Helm, and GitOps; authored runbooks that reduced new-hire onboarding time by 30%.
Cloud Engineer Intern
- Designed and implemented scalable cloud topologies across AWS compute, storage, and networking layers (EC2, S3, RDS).
- Automated infrastructure provisioning and state management utilizing Terraform (IaC).
- Optimized container orchestration and deployment strategies using AWS ECS and Kubernetes (EKS).
Flutter Developer Intern
- Spearheaded the integration of DigiLocker e-KYC utilizing Flutter and OAuth 2.0, significantly streamlining user onboarding.
- Architected and deployed robust modules including dynamic coupon engines, automated GST computation, and secure wallet payment systems.
03. Tech Stack
Languages
Generative AI & LLMs
Backend Architecture
Cloud & DevOps
ArgoCD
Data & Messaging
Observability & Tools
04. Selected Projects
05. Init Handshake
{
"status": "Ready for new opportunities",
"email": "[email protected]",
"availability": "Remote / Full-time / Relocate"
}