Arafat Islam Khan Email & Phone Number

Dhaka, Bd

Team: Reliability Engineering, Experimentation

Dhaka, Bangladesh

Team: Pillars (core infrastructure/platform)- Identified and fixed one of the most severe and complicated bugs in rides vertical, which was a db non-obvious blocking full table scan operation. Collaborated with multiple engineering teams (rides, data-engineering, pillars) to identify the root cause, from dashboards down to exact lines of code. Reduced 30Hour+/year partial outages, also allowing services to have a stricter SLO.- Identified static contents served from object storage, led the integration of edge caching. Edge caching reduced project's total network cost in half, while also dropping end-users latency from 200+ ms to ~30 ms.- Standardized the process of incident analysis, report generation, documentation. Worked with developers to set a healthy blameless postmortem culture and knowledge sharing across all the development teams. Resulted in less repeated similar mistakes from different teams (eg. scan operation on redis), and improving overall production stability.- Integrated and validated open source exporters for self managed redis and elastic-search instances. Used multiple grafana dashboards to gain insights on these self managed instances, resulted in faster time to identify and resolve issues and more overall visibility.- Maintained 20+ self hosted production redis instances, identifying performance issues, scalability issues and life cycle management (update, upgrade, health check, os maintenance, cleanup etc).- Analyzed SRE golden signals to pinpoint issues/abnormalities on 100+ micro-services, propagating to respective service owners. Helped the service owners to mitigate these issues and did followups to ensure services patches are working as intended. - R&D on integration of GitOps and open-telemetry with the existing platform. Building and running demo applications/services with these integration, discussing insights on production adaptation/integration costs and blockers with the lead engineers.

Dhaka, Bangladesh

Team: Backend- Microservice development in GO, building DevOps PaaS on top of Docker & Kubernetes. - Designed and Implemented Connector, a service to handle large volume of IoT events from MQTT broker to securely store or propagate to a air-gapped Kafka cluster, with HA and fault-tolerance.- Contributed to queue service, added latest Kafka versions on the service, also removing old, outdated raw cluster creation logic and moved everything to strimzi (https://strimzi.io) operator. Resulted in faster cluster creation/modification and more stable clusters in the platform.- Contributed to database service, implemented database/cache/queue custom configuration features, so developers can able to create their database/cache/queue with preferred configurations as needed (eg. enable TLS, Replica count, Snapshot TTL).- Designed KloverCloud drive, which acts as an abstraction layer on top of S3, providingusers with a hierarchical filesystem on cloud (SaaS Drive). Broken down the initial monolith design to 10+ microservices, helping lead engineers on the top-level system design process.- Took ownership of the storage service and resolved all the performance related issues. Implemented dynamic threading logic in the service, to list and search/pattern match thousands of files in S3 bucket. Resulted in improved search complexity from polynomial to linear.