I have rearchitect and optimise AWS Glue data-plane storage layer with DynamoDB which enables the service to offer 10X higher concurrent workflow execution limit to the customer. I have optimised DataPipeline infrastructure using AWS IMDSv2 metadata service to reduce the pipeline span up time which reduce cost for customers and implemented CI/CD with automated rollback which also reduce operational effort like on call page. As an L5 engineer I also lead the project.

I am responsible for developing centralised campaigning system where vendor from all over the region can create a promotional campaign for their restaurant or shop and it will be promoted as featured vendor in all product of DeliveryHero. The system was developed using Golang, gRPC and deployed in Kubernetes. Used Vault as config management, Istio as load balancer and proxy, Github Action for CI/CD and Terraform to provision the cluster and pods. After this re design I was able to reduce cost by 23% and onboard 6 child product of DeliveryHero like Foodpanda, Talabat and others. I have orchestrate the service in GKE, integrate with Vault for config management, DataDog to monitoring and GitHub Actions to implement CICD to ship new kind of promotional feature faster and smoother. I was also leading this project of 4 engineers.

I have build centralised order booking system for all of our products using Golang, gRPC and deployed in Kubernetes, GitLab Runner as CI/CD, Grafana to monitor the system and NewRelic for distributed tracing as I also had to oncall like an SRE for the system. I is able to server 300K request per minutes with 90TB of data. GoJek more than 20 product. Previously order was managed by individual production but as we grown it was difficult to maintain individual order booking system for each product and region. So I design and develop highly available order booking system that can server all of our products in all geographical region. We have achieved 99.90% uptime which increase the stability of our platform and reduce the cost about 17%. It also increase our order by 4%. To develop the backend system I used Golang, gRPC, vegeta for load testing, HAProxy, Consul, ScyllaDB, ELK for log aggregation, chef and Grafana for infrastructure related automation and metrics, PagerDuty for alerting and

NewRelic for profiling, Kubernetes, influxDB for time series data store and GCP as Istio cloud provider. Kong API Gateway Platform Engineer at Shohoz Ltd, March 2018 -- Oct 2019 Load Test Shohoz has different types of notification like FCM, APN, SMS, Email. Previously, each product like ride, food, parcel send notification by their own and we didn't have central control of our notification system, billing and failover. So I develop a centralised notification system for our platform. Other product of our platform only can send notification through this system so we have control over all notifications of our platform, retry policy for failover and complete visibility and reporting of billing. It was an event-driven system so I used Kafka for event sourcing, Golang for api, MongoDB, protobuf, gRPC, websocket for communication, ELK for log aggregation, MQTT as message broker, Docker and Kubernetes (ELK) for orchestration and scaling and AWS as cloud provider.

is a video contest website. I was responsible to process HD video into different resolution for better preview. For this I have used AWS Lambda, S3, Glacier, DynamoDB, AWS serverless application model (SAM) framework, Cognito, AWS CodeBuild as pipeline. When ever a file upload to S3 it tiggers lambda function and lambda function will process the video file and store the original file at Glacier as backup. Tea Driver is a private cab booking system. I have developed centralised authentication system using JWT and Golang for entire system. For this I have used Kong as Api gateway and rat limiter. I have also develop our CICD pipeline using GitLab Runner. I also developed uniform logging layer using ELK stack for the entire system, setup alert using Grafana. Also for api gateway log I have steam to Kafka and then store influxDB for postmortem and analysis.

TillBox is business solution with inventory management, tax, accounting and other business services. I have developed the backend system using Spring Boot, MongoDB and MQTT for web service and internal notification system. I have provide clear API documentation using swagger.