Harshit Luthra
Verified Expert in Engineering
DevOps Engineer and Software Developer
Bengaluru, Karnataka, India
Toptal member since August 31, 2021
Harshit is a senior site reliability engineer with 7+ years of experience running cloud native platforms across AWS, GCP, Azure, and on-prem. At TrueFoundry, he built multi-tenant SaaS reliability for enterprise ML workloads, reducing client onboarding by 70% and moving logging to VictoriaLogs, resulting in a 94% drop in query latency. Earlier, he ran infrastructure for a 4 million-user social platform on 95% spot, holding 99.99% uptime while cutting cloud spend by $200,000 a year.
Portfolio
Experience
- Amazon Web Services (AWS) - 7 years
- Kubernetes - 6 years
- Azure Cloud Services - 6 years
- Monitoring - 6 years
- Infrastructure as Code (IaC) - 6 years
- Agentic AI - 5 years
- Google Kubernetes Engine (GKE) - 5 years
- Google Cloud Platform (GCP) - 4 years
Preferred Environment
Kubernetes, Kubernetes Operations (kOps), Google Kubernetes Engine (GKE), Amazon Web Services (AWS), Azure Cloud Services, Terraform, Docker, Prometheus, Helm, Google Cloud Platform (GCP)
The most amazing...
...achievement was keeping a 4 million-user, 1 million-RPM social platform at 99.99% uptime on 95% spot instances while cutting cloud spend by $200,000 a year.
Work Experience
Senior Site Reliability Engineer
TrueFoundry
- Built a modular Terraform framework that cut client onboarding time by 70%, with deployment across AWS, Azure, GCP, and on-prem from minimal config changes.
- Created cloud Marketplace listings (AWS, GCP, Azure) that drove a 40% increase in self-service customer acquisition and reduced sales engineering overhead.
- Migrated the logging stack from Grafana Loki to VictoriaLogs, cutting query latencies by 94%, reducing storage by around 40%, tripling ingestion throughput, and halving CPU/RAM. Published methodology as a public engineering blog.
- Architected severity-tiered incident management integrating Sentry, Grafana, and New Relic with Zenduty and Slack, with team-wise alert routing across five domains, materially reducing MTTR and on-call noise.
- Led platform-wide monitoring overhaul, bifurcating alerts into P0/P1 severity tiers and migrating critical components to New Relic after a successful PoC, improving signal quality while keeping observability spend flat.
- Hardened the platform for multi-tenant SaaS through tighter tenant isolation, namespace-scoped RBAC, and resource governance on shared clusters; standardized labels, annotations, and security contacts across the fleet.
- Served as an escalation point for complex multi-cloud production debugging across EKS, GKE, and AKS, spanning Karpenter autoscaling, EFS/CSI mounts, GPU node scheduling, IAM/IRSA, and airgapped artifact registries.
Infrastructure Lead
Kutumb
- Inaugurated the infrastructure from scratch, including three Amazon virtual private clouds (VPC), private and public subnets, Kubernetes clusters, and VPC peering.
- Used Spot Instances for stateless workloads in Kubernetes using node affinity and cluster autoscaler to reduce infrastructure costs.
- Set up a monitoring and logging stack and alerting tools using the Kubernetes/Prometheus stack, Grafana, Loki, and APM.
- Reduced cloud bills by $200,000 yearly with self-hosting, bin-packing pods, and nodes, Spot nodes, network flow logs check, log, and metrics volume in check.
- Set up and maintained multi-broker Apache Kafka and multi-node ELK Stack on-spot instances.
- Implemented robust network policies and role-based access control (RBAC) configurations, enhancing the security posture of the Kubernetes cluster and reducing potential attack vectors.
- Implemented IPv6 support on the whole infrastructure and added AWS Global Accelerator to reduce latencies in global markets.
- Employed and maintained OpenVPN within the Kubernetes cluster, simplifying networking between on-cluster services and development machines.
- Reduced deployment times to under one minute using caching with self-hosted GitHub Runners on Spot using ArgoCD and Devtron.
DevOps Engineer
Smallcase
- Built AWS SSM to run Ansible playbooks on ASG lifecycle hooks.
- Pioneered a templated CI/CD solution to cater to a multi-tenant environment integration and deployment system using Jenkins.
- Reduced infrastructure costs by moving to better generation instances with a discounted price plan.
Experience
CI/CD Project Using CircleCI
https://github.com/sachincool/AutoDeploy-SuperPower-ProjectInfrastructure Lead
https://crafto.appTo maintain optimal system health and facilitate rapid issue resolution, I incorporated the principles of RED (rate, errors, duration) and USE (utilization, saturation, errors) monitoring. These practices provide comprehensive insights into the cluster's performance and any potential issues, thus enabling teams to identify and rectify any operational anomalies promptly.
To further enhance the agility and productivity of our development teams, I implemented ArgoCD for continuous deployment. This approach abstracts the complexities of deployment processes, allowing developers to remain focused on critical business requirements. As a result, they can deliver features and fixes more rapidly and reliably, which ultimately contributes to our organization's competitive advantage.
Kafka Debezium CDC Pipeline
DIAGRAM
MySQL tables | Kafka connector | Kafka topics | Kafka connector | Elasticsearch index
I added proactive monitoring and alerting.
IPv6 and Global Accelerator for Global Markets
I monitored VPC flow logs and integrated ALB logs with Amazon S3 bucket
and prepared Athena tables and queries.
Education
Bachelor of Engineering Degree in Computer Science and Engineering
Chitkara University - Punjab, India
Certifications
Claude Certified Architect – Foundations
Anthropic
Certified Kubernetes Security Specialist
Cloud Native Computing Foundation
Certified Kubernetes Administrator
CNCF
AWS Certified Solutions Architect Associate
AWS
Skills
Libraries/APIs
Amazon API, Terragrunt, Node.js, vLLM, PySpark, Claude API
Tools
Grafana, Helm, Amazon CloudWatch, Terraform, AWS Fargate, Observability Tools, AWS IAM, Amazon OpenSearch, Google Kubernetes Engine (GKE), Jenkins, Amazon Elastic Container Service (ECS), GitLab CI/CD, Amazon ElastiCache, Grafana k6, Amazon SageMaker, Istio, Sentry, VPN, CircleCI, Ansible, AWS CloudFormation, Kafka Streams, Amazon EKS, AWS ELB, Kubernetes Operators, GCP Security, OpenTofu, RabbitMQ, Amazon Athena, AI Prompts, Claude Code, Claude Agent SDK, Claude
Paradigms
DevOps, Load Testing, Automation, Serverless Architecture, Continuous Delivery (CD), Continuous Integration (CI), Azure DevOps
Platforms
Kubernetes, AWS Lambda, Amazon Web Services (AWS), Amazon EC2, Docker, Google Cloud Platform (GCP), Apache Kafka, Linux, Vanta, New Relic, Langfuse, Confluent, Azure, Kubeflow
Storage
Amazon DynamoDB, Elasticsearch, Database Administration (DBA), Databases, Azure Cloud Services, Redis, Cassandra, MongoDB, MySQL, PostgreSQL, Druid.io
Languages
Bash, Python 3, Python, Go, C++11
Frameworks
AWS HA
Industry Expertise
Cybersecurity
Other
Kubernetes Operations (kOps), Amazon RDS, CI/CD Pipelines, Containerization, GitHub Actions, Machine Learning Operations (MLOps), Site Reliability Engineering (SRE), Cloud, Distributed Systems, Infrastructure as Code (IaC), Argo CD, Networking, GPU Computing, Graphics Processing Unit (GPU), Identity & Access Management (IAM), Cloudflare, Agentic AI, AI Automation, Release Management, Observability, AWS Cloud Security, FedRAMP, Amazon Bedrock AgentCore, OpenTelemetry, Apache Cassandra, AWS DevOps, AWS Certified DevOps Engineer, Cloud Infrastructure, Message Queues, Firewalls, AWS Certified Solution Architect, Amazon API Gateway, AWS VPN, Queue Management, Cloud Architecture, Large Language Model Operations (LLMOps), Large Language Models (LLMs), Incident Management, Infrastructure Testing, SOC 2, System Administration, GCP DevOps, Capacity Planning, VictoriaLogs, Karpenter, Artificial Intelligence (AI), QA Automation, Quality Assurance (QA), Software, FastAPI, Monitoring, Structured Logging, Operating Systems, Computer Networking, Prometheus, Lambda Functions, Relational Database Services (RDS), AI Architecture, MLflow, AI Model Training, Cloud Computing, Firmware, Claude Design, Prompt Engineering
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring