Samarth Narula

Description: We are pioneers into K8 based Multi Tenant Automation. You must have heard Virtualisation & Namespaces, what we have done is many levels above that

​

Project : Silo Multi-Tenant Model Setup on Customer AWS Account Automation
Position: Principal Cloud Engineer
Tech Stack: AWS Services like Cloud formation, SNS, S3, EKS, EC2, Resource Groups, Lambda Functions, System Manager Documents then Parameter Store, Secret Manager, RDS, Cloudfront, VPC and then Kubernetes, Docker, Helm Charts, Argo CD and Agro Workflows, Github Actions, Java, Python and Polyglot Programming.

Work Description : Architected and Implemented the Automation. Customers will use a cloud formation template, which will spin all the infra required for the product and also generate kubeconfig file and values.yaml file(with customer public info) and place by executing k8 job while cf template execution and place it into S3 bucket and then cf template will also create a lambda function which will have S3 create event trigger, which will generate a pre-signed URL for the k8 config amd values.yaml file and send it to our public SNS topic and in our account I have written another Lambda which will process this pre-signed url and store the files received data in secret manager and RDS postgres aurora database and then this lambda will trigger a multi-step Argo Workflow:
Step:1 Add customer cluster in our Loft VCluster instance.
Step 2: Create Vcluster in Physical VCluster
Step 3: Update cloudfront orgin in customer account access through k8 job using k8 service account which has assumed a customer aws account role with required permissions.
Step 4:Create parameter store in customer parameter store, copies parameters from our parameter store to customer parameter store.
Step 5:create literal file for registering app on our ArgoCD.
Step 6:Deploy Atlan under VCluster: Creates app in ArgoCd and lets it sync the helm charts changes and deploy atlan product
All of the above completed in 74 days of joining the company

After the above work completed, now

1. Enhancing the Customer AWS Multi-Tenant Setup, builded the tenant off-boarding workflow, figuring out solutions for all tenants off-boarding workflow. 

2. Helping in identifying and finding solutions for AWS Accounts Cost Optimisation.

3. Helping new joiners in project to get upto Speed and creating training material for new joiners
    

Project : Existing Multi-Tenant Setup
Description: Deploy Atlan instance to our AWS EKS Cluster VCluster using Loft VCluster, ArgoCD, Argo Workflows, Helm Charts and for IAAC Terraform

​

Project: AWS Cost Optimisation
Role : Solution Architect
1. Identified key services in use causing high cost using AWS Quick Sight reports.
2. Connected with couple of Senior Startup Solution Architect guys from aws company to get more ideas and also attended aws cost optimisation seminars to learn more techniques.
2. Promptly brought down cost by 800$/day by disabling dev environment EKS CloudWatch Logs.
3. Recommended the below technics to further bring down the overall cost :
* Update existing eks nodes to use AMD series Processor across all environments.
* Move all OnDemand instances to all spot for dev Environment Instances and One Ondemand and rest spot for Production via Terraform Script as Terraform was used for spinning all the infra, if we do directly from aws console it will lead to terraform state file drift. As the state was stores in S3 bucket.
* To enable logging back in dev environment reduce the logging period and configure lifecycle rules for moving data to glacier storage.
* For Lambda use ARM based architecture, migrate all existing *86 Architectures to ARM based Architectures
4. Assigned these task to other engineers in team and ensured proper execution.

RND projects:
1) POC: Send Cross Account AWS SNS Topic Notification from Lambda Function.
2) POC: Test and check if we can send cross account sns notification from AWS CloudFormation using AWS CLI, it is possible by using CLI but not through AWS Console.
Outcome: Raised a AWS Feature request on AWS Cloud Formation RoadMap GitHub Repo to enable sending cross account SNS topic notification using AWS Console.
All of the above completed in 15 days

​

Project: Create Release Management RoadMap for the Company
Role: Project Owner & Enterprise Architect | One Man Army Role.
Work Done:
1. Identified Problems with current release processes.
2. Explored different solutions in market which can be used as Docker Image and Helm Chart Package Repository.
3. Listed pros and cons for different registry solutions and cost for each Solution.
4. Finally proposed solutions for Helm Chart Repository & Docker Repository, Plugins and architectural changes to enable proper Semantic Number Versioning of Artefacts in automated fashion by using git action and Sever side git hook and Client side hook in automated fashion by using Husky (NPM Package) and installing git cz plugin. 5. Also Recommended the future enhancements like since we use Github, automated release documentation with Github Pages, Enable Github Community for long term growing collaboration for engineers. 6. While creating this roadmap compared two data transfer cost effective solutions.
5. Then lead a team of 2 engineers in implementing the proposed architecture during the initial few task, then alone everything
6. Created various estimated cost reports using AWS Calculator.
7) Made and executed critical decisions around Private ECR(no caching support), VPC NAT vs VPC PrivateLink connection with EKS for imagepull, Cross Region And Cross Account Replication strategy.
8) Wrote Github Action Workflow, Scripts for CICD, push image to ECR, pull latest image from ECR, Create ECR repo if not exist, Create Lifecycle policy if not exist for ECR, keep last 10 github tags, delete all tag on feature branch deletion. Semantic Versioning.
11) Design Strategies with principal of least privilege around ImagePull in eks from ecr, Image push from github workflows using IAM OIDC Roles.
12) Design and Architect Strategies for rollback mechanism, deployment strategies like blue green, Canary, rolling updates, Caching, Replication. Automated Image Update with ArgoCD image updater etc.

​

Project: Create Loft VCluster Course
Role: Content Creator and Course Instructor
Work Done:
1. Created Curriculum and Content for the course.
2. Created Enterprise level Spring Boot Application.
3. Created kubernetes deployment and load balancer type service yaml files for the spring boot application.
4. Created AWS EKS Cluster and deployed spring boot application into it.
5. Created Helm Chart for Spring Boot application.
6. Deployed loft VCluster into AWS EKS and created VCluster and deployed Spring Boot application into it.

Other Architectural Discussions:
1. Interactions with Google Employees and our vendor partner for POC on Google Anthos.
2. Interactions with junior engineers working on different cutting edge new technology initiatives.
3. Interactions with Amazon software engineers and solutions architects for solving specific architectural problems.

​

General Roles & Responsibilities:
1) Planning engineering strategies for a company
2) Implementing process improvements
3) Managing engineering departments in tasks like research and design
4) Providing expert advice to other engineers
5) Determining department goals and creating implementation plans
6) Creating and managing engineering budgets around estimated AWS service usage, Github Plans and other toolings.
7) Implementing process improvements
8) Architecting, Planning, Managing & also Executing when other junior engineers are not available.

Exposure and walkthrough of new technology initiatives around SRE, Security and Automation.
Exposure to kubecost, teleport, Datadog, Rootly etc

​

​