SaaS / B2B Software · Adtech Platform Devops Foundation
groweasy
Multi-tenant AdTech SaaS Platform.
The brief
Deliver a scalable, production-ready AdTech SaaS platform with a React-based frontend (including Creative AI and Rules Engine modules), migrate backend microservices from Elastic Beanstalk to Kubernetes on AWS EKS, and establish a full DevOps foundation (CI/CD, observability, security) capable of supporting a growing multi-tenant customer base across Google, Amazon, and Facebook advertising channels.
What we built
A production-grade AdTech platform ("Maino/Manthan") for managing and optimising advertising campaigns across Google Ads, Amazon Ads, and Facebook. The platform includes:
- a React.js frontend with a full design system, Creative AI module (ad performance analysis, ML-driven recommendations, DoD graphs), Rules Engine with Firebase-backed state, Keyword Manager, and Real-time Reports
- a microservices backend (bff, maino_analytics, orchestrator, reports-manager, celery, google-ads-insights, creative-studio, rules-engine, and others) on Kubernetes (AWS EKS)
- full observability via EFK stack + Prometheus/Grafana/Alertmanager
- event streaming via Kafka (MSK), RabbitMQ, and SQS
- a data/ML layer with SageMaker, Redshift, RStudio, JupyterHub, and Databricks; and
- SOC2 compliance infrastructure via Sprinto. Infra was migrated from Elastic Beanstalk to EKS over the engagement and continuously optimised for cost and availability.
Production Kubernetes cluster on AWS EKS running 15+ microservices with automated CI/CD (Jenkins multibranch pipelines), EFK logging, Prometheus/Grafana/Alertmanager monitoring, Kafka/MSK event streaming, RDS (multi-AZ with read replicas), Redshift, SageMaker, RStudio/JupyterHub, WordPress (content/blog), and SOC2 compliance infrastructure via Sprinto. React frontend with Creative AI, Rules Engine, and Real-time Reports modules deployed. As of mid-2025, a cluster consolidation (prod→stage) cost optimisation is underway.
Delivery timeline
How it was built, phase by phase.
8 workstreams across 83 weeks of operated delivery.
- deployDec 2023 – Jan 2024
Production Kubernetes Infrastructure Setup
End-to-end setup of AWS EKS clusters for both staging and production environments, including ingress controllers, node groups, namespaces, configmaps, and deployment manifests for a large microservices fleet.
Multiple EKS clusters (staging and production) operational with ingress, TLS, domain routing.
AWS EKSKubernetesHelmkubectlk9sLens - buildDec 2023 – Jun 2025
CI/CD Pipeline Engineering (Jenkins)
Design and continuous improvement of Jenkins-based CI/CD pipelines for 20+ microservices, including multibranch pipelines, parameterised builds, GitHub webhook integration, Docker build optimisation.
Fully automated CI/CD covering staging and production for all microservices; zero-downtime rolling deployments with GitHub commit ID traceability.
JenkinsGitHub WebhooksDockerMultibranch PipelineArgoCD (explored)AWS ECR - stabilizeDec 2023 – Jun 2025
Observability Stack (EFK + Prometheus/Grafana)
Full observability layer deployment: EFK (Elasticsearch-Fluentd-Kibana) for container log aggregation, Prometheus + Grafana for metrics, Alertmanager + Slack/SNS for alerting, and Uptime Kuma for endpoint monitoring.
Full-stack observability live on both staging and prod: log aggregation with rotation policies, custom Grafana dashboards.
ElasticsearchFluentdKibanaPrometheusGrafanaAlertmanager - deployDec 2023 – Jun 2025
Microservices Fleet Deployment & Management
Continuous deployment and lifecycle management of 15+ named microservices including maino, manthan, bff-revamp, celery, orchestrator, reports-manager, google-ads-insights, creative-studio, rules-engine, anushudan.
Entire AdTech platform microservice fleet running on Kubernetes with individual CI/CD pipelines.
KubernetesAWS EKSDockerJenkinsAWS RDSAWS Lambda - integrateDec 2023 – Jul 2024
Message Queue & Event Streaming Infrastructure
Deployment and integration of RabbitMQ, Apache Kafka (self-hosted on K8s and AWS MSK), and SQS as the messaging backbone for async inter-service communication and ad-platform event pipelines.
Multi-broker messaging layer operational: RabbitMQ for task queuing, Kafka (K8s-native then migrated to AWS MSK) for event streaming.
RabbitMQApache KafkaAWS MSKAWS SQSVPC ConfigKubernetes - stabilizeApr 2024 – Nov 2024
Security Hardening & SOC2 Compliance
Systematic AWS security hardening and SOC2 compliance readiness via Sprinto: IAM policies, VPC security groups, NACL reviews, S3 KMS encryption, CloudTrail, GuardDuty, MFA, VPC flow logs, ECR encryption.
SOC2 compliance checklist substantially completed via Sprinto; all S3 buckets encrypted, RDS moved to private subnets with VPC peering.
AWS IAMAWS GuardDutyAWS CloudTrailAWS KMSVPC Flow LogsSprinto - buildJan 2024 – Mar 2024
React Frontend (Creative AI & Ad Management UI)
React.js frontend rebuild for an AdTech dashboard: design system from scratch (components, theming, MobX state), Creative AI module (performance insights, recommendation flows, ad creative analysis).
Full design system built from scratch; Creative AI module with performance graphs, recommendation flows, DoD views, and outlier detection.
React JSMobXFirebase/FirestoreCSS ThemingNext.js (middleware/private routing)Facebook Auth - integrateApr 2024 – Jun 2025
Data & ML Infrastructure (SageMaker, Redshift, RStudio, JupyterHub)
Setup and management of AWS SageMaker notebook instances, Redshift data warehouse, RStudio/JupyterHub multi-user data science environments.
Multi-user data science workspace operational: SageMaker for model training, Redshift for warehousing ad analytics data.
AWS SageMakerAWS RedshiftRStudioJupyterHubDatabricksAWS S3
More case studies
Related work
09 · Run a function
Stop renting hours. Start running functions.
Pick the function you want off your plate. We'll map the brain and name the outcome we'd commit to — before you do.