Mustafa – Ansible, Kubernetes, Python

AI infrastructure and platform powering compute, orchestration, training, simulation, and inference for next-generation physical AI humanoid robots.

• Built and led AI infrastructure and platform serving as the compute and orchestration backbone for next-generation physical AI robots.
• Designed and operated multi-cluster Kubernetes platforms across EKS, k3s, and kubeadm environments.
• Designed and operated GPU-accelerated AI infrastructure with hundreds of Nvidia GPUs, including H100, L40s, and A10Gs.
• Implemented GPU scheduling, resource isolation, quotas, and GPU-aware autoscaling.
• Implemented Nvidia MIG partitioning to improve cost efficiency, utilization, and availability.
• Delivered object-storage-backed and high-performance storage solutions using EFS and integrated S3 for cost-efficient data access.
• Built persistent developer environments on Kubernetes to support experimentation and interactive debugging near data and GPUs.
• Implemented multi-node training orchestration with Volcano and Ray.
• Built end-to-end training and inference pipelines and integrated them with CI/CD and infrastructure as code.
• Improved observability for AI workloads by monitoring GPU health, data pipelines, and model serving behavior.
• Scaled the platform foundation for reliability and cost efficiency as the organization grew.

An autonomous trading desk that uses automated trading strategies to analyze market conditions and execute trades with minimal manual intervention. The platform supports systematic trading operations, aiming to improve execution efficiency and enable consistent, data-driven trading decisions.

• Delivered a new algo trading desk where automatic trading decisions were executed from the combined output of traditional strategies, real-time data, sentiment analysis, and AI models trained on historical stock data.
• Built data infrastructure and robust pipelines to process and store large-scale real-time streaming data.
• Implemented quantitative models using AI training, sentiment analysis, and traditional trading strategies.
• Architected and implemented infrastructure, libraries, and services to run algo trading pipelines and algorithms.
• Used Telegram for trading notifications.

An observability platform running on an Amazon EKS Kubernetes cluster, providing application monitoring and visibility through the Grafana OSS suite. It incorporates Grafana Faro and Grafana OnCall to expand application observability and incident response capabilities while reducing costs through open-source tools. Infrastructure components are provisioned and managed using Terraform and GitOps principles, ensuring consistent and immutable infrastructure.

• Delivered and maintained monitoring services on an EKS Kubernetes cluster using the Grafana OSS suite.
• Implemented new features to increase application visibility and reduce costs through open-source tooling.
• Achieved immutable infrastructure by provisioning infrastructure components with Terraform and GitOps principles.

An observability platform running on an Amazon EKS Kubernetes cluster and powered by the Grafana OSS suite. It enhances application visibility and incident response through Grafana Faro and Grafana OnCall, while reducing operational costs through open-source tools. All infrastructure components are provisioned and managed using Terraform and GitOps principles to ensure consistent, immutable infrastructure.

• Delivered microservices running on Kubernetes clusters across AWS regions with a focus on the EKS platform.
• Maintained EKS clusters by regularly applying version upgrades in collaboration with service owners.
• Supported developers in promoting changes into production securely by automating release-cycle stages.

A cloud-native platform focused on security, observability, and automated infrastructure management. It strengthens the existing platform through automated secret management, Web Application Firewall integration, vulnerability analysis, and file integrity monitoring. Self-service developer platforms embed security into earlier stages of the development lifecycle, enabling teams to detect defects sooner. Cloud-native microservices benefit from enhanced observability, while infrastructure and platforms across multiple regions, from the US to the Far East, are managed as code and deployed through fully automated pipelines.

• Delivered security remediation and layered defense into the existing platform by automating secret management, web application firewall integration, vulnerability analysis, and file integrity monitoring.
• Enabled self-service platforms for developers and shifted security measures left to earlier delivery stages.
• Increased observability across the cloud-native microservices network.
• Implemented infrastructure and platforms across multiple regions as code and promoted them through fully automated pipelines.