Kaleb Clement Sembiring

Global E-commerce Service Architecture Platform

Bytedance, Jakarta

February 2024 - PRESENT

Serve as the primary maintainer for 14 core middleware platforms (including storage, caching, and scheduling), delivering 99.99% reliability by leading 24/7 on-call operations and major cross-functional incident responses for Global E-commerce teams distributed across the US, Asia, Europe, and the Rest of the World.
Built an AI-powered internal assistant using custom language models (Openclaw) to automate root-cause analysis for common system incidents, reducing engineering workload by 50-70% and cutting total triage time by nearly 70%.
Enhanced operational stability and monitoring for large-scale job scheduling systems, eliminating approximately 100,000 monthly failures and reducing incident recovery times (MTTR) by 25%.
Drove company-wide engineering initiatives by standardizing service agreements (SLAs), building real-time stability dashboards, and streamlining incident management to cut QA triage time by 40%.
Orchestrated the zero-downtime migration of 60+ Tokopedia services to ByteDance's internal cloud by building a custom orchestration platform across GCP, AWS, and Alibaba Cloud.

Automated First Responder: Engineered a custom AI bot utilizing Lark Bot and Openclaw to autonomously manage and monitor 14 core middleware platforms. Programmed the bot to seamlessly intervene and initiate diagnostics if an on-call engineer does not respond within 5 minutes of a mention.
Dynamic Infrastructure Querying: Developed advanced capabilities allowing the AI to securely authenticate via ByteCloud service accounts to dynamically query service monitoring systems, extract infrastructure metrics, and parse system logs in real-time.
Accelerated Triage: Leveraged custom language models to autonomously analyze incident data and perform root-cause analysis, successfully reducing incident triage time by nearly 70% and significantly cutting overall engineering workload.

Cloud Platform Engineer

Tokopedia, Jakarta

August 2021 - February 2024

Led the enterprise-wide migration from Jenkins to GitHub Actions, standardizing CI/CD pipelines (including Ansible, Packer, and Terraform) across Tokopedia and its subsidiaries.
Enabled engineers to develop CI/CD use cases using self-hosted GitHub Runners, significantly enhancing pipeline scalability and maintainability.
Automated the identification and decommissioning of over-provisioned infrastructure, driving significant cost savings and improving overall resource utilization.
Handled daily infrastructure operations and performance tuning across GCP, AWS, and Alibaba Cloud. Collaborated with the GoTo-Financial team to successfully implement and migrate Kubernetes for GoTo-Logistic companies.

Scalable CI/CD Architecture: Architected enterprise-wide GitHub Actions CI/CD pipelines utilizing self-hosted Kubernetes runners and custom Helm charts to enable dynamic Horizontal Pod Autoscaling (HPA).
Standardized Tooling: Empowered global engineering teams by deploying organizational-level custom runner images pre-configured with Ansible, Docker, Packer, and Consul, seamlessly connecting staging and beta environments for performance testing.
GitOps & Deployment: Implemented automated GitOps workflows utilizing self-testing loops and pull request (PR) comment-based deployment approvals.
Multi-Cloud Image Provisioning: Streamlined infrastructure deployment by automating image builds (via Packer and Ansible) to securely deploy pre-configured application components across GCP, AWS, Alibaba Cloud, and Azure for compute instance provisioning (e.g., EC2, GCE).

Backend Engineer

Tricada Intronik, Bandung

April 2021 - August 2021

Created and maintained service applications, designing robust product APIs between services for frontend and UI/UX teams according to client requirements. Designed and executed queries for flow processing jobs using KSQL and Kafka.
Deployed and monitored applications automatically using Jenkins, Kubernetes, and other deployment tools. Implemented comprehensive alert and notification systems using Grafana, Prometheus, and scheduling.