Devashish Gupta

About Devashish Gupta

Strategic Senior DevOps Engineer with a proven track record of architecting mission-critical, scalable, and secure cloud infrastructure for high-growth AI startups. An expert in translating complex business requirements into robust technical solutions that enhance performance, ensure reliability, and control costs.Core Competencies:• MLOps & AI Infrastructure: Designing and deploying high-density model serving platforms (EKS, Triton, ModelMesh) and event-driven batch processing pipelines (AWS Batch, Step Functions).• Cloud Architecture & Resilience: Architecting multi-region disaster recovery strategies, database optimization patterns, and highly available, cost-efficient AWS environments.• DevSecOps & Automation: Implementing \"Shift Left\" security in CI/CD, automating developer workflows (GitHub Bots), and leading security analysis and remediation (AWS GuardDuty).• Governance & Compliance: Driving initiatives to reduce technical debt, implement cloud governance (AWS Control Tower), and ensure adherence to standards like SOC 2 and ISO 27001.GitHub : https://github.com/dcgmechanicsMedium : https://dcgmechanics.medium.com- Updated On 16 Nov 2025 -

Experience

1. Devops Engineer 2

   Spyne

   May 2025 — Present · Gurugram, IN

   Cloud Architecture & Performance Optimisation:• Resolved RDS performance bottlenecks by architecting a \"Cache as a Gatekeeper\" pattern (Redis), an Aurora migration strategy, and SQS decoupling for high-volume writes.• Mitigated L7 DDoS attacks by implementing an AWS WAF rate-limiting solution, successfully handling traffic spikes exceeding requests per minute.• Designed ECS Service Connect to reduce latency, cut data transfer costs, and secure internal service communication.AI/ML Platform Engineering (MLOps):• Architected a scalable, event-driven ML batch processing pipeline (SQS, Step Functions, AWS Batch) for asynchronous model training/inference.• Designed a high-density MLOps platform on AWS EKS using ModelMesh and NVIDIA Triton to efficiently serve hundred of multi-modal AI models.• Developed scaling strategies for ML workloads, defining principles for \"warm pools\" Spot Instance usage for up to 90% cost savings.Automation, Security & Disaster Recovery:• Engineered a GitHub App (Bot) from scratch to automate JIRA validation within pull requests, enforcing development best practices and blocking unlinked merges.• Designed a comprehensive, multi-region \"Warm Standby\" disaster recovery (DR) plan for all critical AWS services (ECS, RDS, SQS) to ensure business continuity.• Conducted security analysis with AWS GuardDuty, Implementing remediation plans for high-severity risks like data exfiltration and exposed resources.• Authored a strategic roadmap to reduce technical debt, proposing migrations from Jenkins to AWS CodePipeline and adopting AWS Control Tower for a multi-account strategy.Cost Optimisation & Cloud Migration:• Led complex cloud migrations, including moving a memory-intensive database to GCP using Terraform and strategically migrating all AI workloads from GCP back to AWS.• Achieved significant cost savings by migrating GPU instances to more cost-effective types (g5 to g6) and downsizing under-utilised Redis & RDS instances.

Education

• Maharshi Dayanand University

  Bachelor of Technology - BTech, Computer Software Engineering

• College of Commerce

  Higher Secondary , Science