SaiKumar M profile photo 1 on Tal

SaiKumar M

AI Platform Engineer @ Delta | MLOps & GenAI Infra | RAG | AWS Pro | MCP | Agentic Foundations

Bengaluru

7+ years of work experience

Delta Air Lines
Dec 20241 yr 10 mosPresent

Bengaluru

• Leading Next-Gen multi-tenant Kubernetes platform migration on OpenShift (ROSA), reducing compute spend by 30% for 500+ production services • Building internal developer platform capabilities enabling self-service provisioning, secret management & deployment automation for 100+ teams • Rolled out GitOps with GitLab CI + Helm for Day-0/Day-1 automation, cutting onboarding time from 5 days → < 1 hour • Modernizing observability using OpenTelemetry pipelines → 40% faster RCA and improved incident response • Architected a high-throughput enterprise RAG pipeline using AWS Bedrock and custom MCP servers, reducing document intelligence latency by 42% and securing agentic data retrieval. • Enhancing SRE and DevEx by implementing progressive delivery, proactive health checks & automated remediation Tech stack: OpenShift (ROSA), Kubernetes, GitLab CI, Helm, AWS CDK, CloudFormation, ELK, OpenTelemetry, API Gateway, Serverless
Carelon Global Solutions
Jan 20232 yrsDec 2024

Bengaluru, Karnataka, India

• Automated infrastructure provisioning using Terraform & Helm → 50% faster deployments and zero config drift across clouds (AWS & Azure) • Managed large-scale microservices on EKS & AKS with GitOps (ArgoCD) → 50% faster release cycle and improved scalability • Optimized Kafka & Spark workloads in Kubernetes achieving 99.95% uptime while supporting 50% higher data throughput • Built cloud access automation using Python to remove 80% of manual effort for 500+ engineers • Reduced compute wastage by 20% using KEDA auto-scaling and improved cost efficiency • Enhanced security posture by implementing Teleport for identity-based secure access • Designed self-healing node lifecycle automation using Karpenter & Infra-as-Code → higher availability & lower operational toil • Integrated Kubeflow & Crossplane to enable early MLOps adoption and cross-cloud infrastructure abstractions • Standardized Observability stack with Splunk & OpenTelemetry, cutting MTTR by 35% Tech & Tools: AWS, Azure, Kubernetes (EKS/AKS), Terraform, Helm, ArgoCD, KEDA, Karpenter, Python, Teleport, Splunk, Istio, GitOps, Crossplane, Kubeflow
NexInfo Solutions, Inc.
Jun 20193 yrs 7 mosDec 2022

• Engineered end-to-end observability using Prometheus, Grafana & ELK → 50% faster incident response and improved system reliability • Designed production-grade CI/CD with Azure DevOps + ArgoCD → 60% higher deployment frequency and near zero-downtime rollouts • Improved message resilience by integrating RabbitMQ + Kyverno policies → <5% message failure rate across microservices • • Designed and managed cloud infrastructure on Google Cloud Platform (GCP) including GKE, Cloud Storage, Pub/Sub, and BigQuery to support scalable microservices. • Automated monthly patching using Ansible → 25% reduction in patching effort + consistent compliance • Led 24/7 production support with 30-minute average issue resolution and deep RCA to eliminate repeat failures • Strengthened cloud reliability with Kubernetes (EKS/AKS) based scaling and proactive capacity monitoring Tech Stack: Kubernetes (EKS/AKS), Azure DevOps, ArgoCD, Terraform, Ansible, Prometheus, Grafana, ELK, RabbitMQ, Kyverno, Jenkins, Docker, Bitbucket

SaiKumar’s AI Stack

cursor
cursor
github copilot
github copilot
langchain
langchain
ollama
ollama

Education

Sri sairam engineering college

Mechanical Engineering

Sri sairam engineering college

Join SaiKumar and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

SaiKumar M on Tal

Headline

AI Platform Engineer @ Delta | MLOps & GenAI Infra | RAG | AWS Pro | MCP | Agentic Foundations

About

I build the nervous systems for modern enterprise scale. My engineering philosophy revolves around a singular truth: the true value of Generative AI cannot be unlocked without a highly resilient, distributed, and deterministic cloud fabric beneath it. I specialize in the intersection of hard-core distributed systems and cutting-edge LLMOps specifically operationalizing autonomous AI agents to perform complex, stateful tasks at enterprise scale.My journey began by engineering advanced multi-cloud observability and high-availability frameworks at Nexinfo Solutions, which paved the way for spearheading massive distributed systems migrations at Carelon Global Solutions. Today, as a Senior Platform Engineer at Delta Air Lines, I fuse this deep infrastructure expertise with modern AI orchestration. I architect production-grade GenAI environments and multi-cloud AI gateways that power intelligent, autonomous enterprise applications. 🚀 Core Technical Pillars: GenAI & Agentic Systems: AWS Bedrock, Azure OpenAI Service, LangGraph, AgentCore, Model Context Protocol (MCP), MLflow, LLM Model Serving.Cloud Fabric & Platform Engineering: Kubernetes (EKS/ROSA), Infrastructure as Code (Terraform), GitOps (ArgoCD, GitLab CI), Enterprise Observability (Prometheus, Grafana, Datadog). Credentials: Certified Kubernetes Administrator (CKA) | AWS Certified Solutions Architect – Professional.With nearly 6 years of hands-on platform engineering experience, I am currently exploring hybrid opportunities within Bengaluru’s elite tier-1 tech ecosystem. I am available on a short notice period and ready to architect your next AI-native infrastructure leap. Let's build

Location

Bengaluru

AI stack

cursor

github copilot

langchain

ollama

Work experience

Senior Platform Engineer - AI Infra at Delta Air Lines, Dec 2024 - Present · 1 yr 10 mos, Bengaluru, • Leading Next-Gen multi-tenant Kubernetes platform migration on OpenShift (ROSA), reducing compute spend by 30% for 500+ production services • Building internal developer platform capabilities enabling self-service provisioning, secret management & deployment automation for 100+ teams • Rolled out GitOps with GitLab CI + Helm for Day-0/Day-1 automation, cutting onboarding time from 5 days → < 1 hour • Modernizing observability using OpenTelemetry pipelines → 40% faster RCA and improved incident response • Architected a high-throughput enterprise RAG pipeline using AWS Bedrock and custom MCP servers, reducing document intelligence latency by 42% and securing agentic data retrieval. • Enhancing SRE and DevEx by implementing progressive delivery, proactive health checks & automated remediation Tech stack: OpenShift (ROSA), Kubernetes, GitLab CI, Helm, AWS CDK, CloudFormation, ELK, OpenTelemetry, API Gateway, Serverless

DevOps Engineer at Carelon Global Solutions, Jan 2023 - Dec 2024 · 2 yrs, Bengaluru, Karnataka, India, • Automated infrastructure provisioning using Terraform & Helm → 50% faster deployments and zero config drift across clouds (AWS & Azure) • Managed large-scale microservices on EKS & AKS with GitOps (ArgoCD) → 50% faster release cycle and improved scalability • Optimized Kafka & Spark workloads in Kubernetes achieving 99.95% uptime while supporting 50% higher data throughput • Built cloud access automation using Python to remove 80% of manual effort for 500+ engineers • Reduced compute wastage by 20% using KEDA auto-scaling and improved cost efficiency • Enhanced security posture by implementing Teleport for identity-based secure access • Designed self-healing node lifecycle automation using Karpenter & Infra-as-Code → higher availability & lower operational toil • Integrated Kubeflow & Crossplane to enable early MLOps adoption and cross-cloud infrastructure abstractions • Standardized Observability stack with Splunk & OpenTelemetry, cutting MTTR by 35% Tech & Tools: AWS, Azure, Kubernetes (EKS/AKS), Terraform, Helm, ArgoCD, KEDA, Karpenter, Python, Teleport, Splunk, Istio, GitOps, Crossplane, Kubeflow

DevOps SRE Engineer at NexInfo Solutions, Inc., Jun 2019 - Dec 2022 · 3 yrs 7 mos, • Engineered end-to-end observability using Prometheus, Grafana & ELK → 50% faster incident response and improved system reliability • Designed production-grade CI/CD with Azure DevOps + ArgoCD → 60% higher deployment frequency and near zero-downtime rollouts • Improved message resilience by integrating RabbitMQ + Kyverno policies → <5% message failure rate across microservices • • Designed and managed cloud infrastructure on Google Cloud Platform (GCP) including GKE, Cloud Storage, Pub/Sub, and BigQuery to support scalable microservices. • Automated monthly patching using Ansible → 25% reduction in patching effort + consistent compliance • Led 24/7 production support with 30-minute average issue resolution and deep RCA to eliminate repeat failures • Strengthened cloud reliability with Kubernetes (EKS/AKS) based scaling and proactive capacity monitoring Tech Stack: Kubernetes (EKS/AKS), Azure DevOps, ArgoCD, Terraform, Ansible, Prometheus, Grafana, ELK, RabbitMQ, Kyverno, Jenkins, Docker, Bitbucket

Education

Mechanical Engineering, Sri sairam engineering college