Maaz is available for hire
Hire MaazMaaz Khalid
Verified Expert in Engineering
AI Infrastructure Architect and Developer
Abu Dhabi, United Arab Emirates
Toptal member since October 12, 2026
Bio
Maaz is an AI infrastructure architect with 10+ years of experience with OpenStack, Ceph, and Kubernetes. He builds GPU platforms ranging from bare-metal to production LLM inference, including air-gapped deployments. Maaz designed AI Factory architectures for Core42 (G42) with NVIDIA and AMD GPUs on rail-optimized InfiniBand, and architected a sovereign AI Token Factory optimized with NVIDIA Dynamo and vLLM.
Portfolio
Net Group
Kubernetes, NVIDIA CUDA, AMD, InfiniBand, OpenAPI, vLLM, Packer, MAAS, Go...
Coredge.io
OpenStack, Ceph, MAAS, Jenkins, Ansible, Grafana, Prometheus, OpenShift, Loki...
Snapdeal
OpenStack, Ceph, Kubernetes, Jenkins, Ansible, Wazuh, Loki, Helm...
Experience
- OpenStack - 8 years
- Ceph - 8 years
- Kubernetes - 7 years
- Large Language Models (LLMs) - 3 years
- GPU Clusters - 3 years
- NVIDIA CUDA - 2 years
- Key-Value (KV) Cache - 2 years
- DynamoAI - 1 year
Work Experience
Senior Cloud Architect
2025 - PRESENT
Net Group
- Architected the company's AI cloud platform, delivering Kubernetes-as-a-Service (KasS) and Bare-Metal-as-a-Service (BMaaS) on NVIDIA and AMD GPUs as a self-service product.
- Designed the AI cloud solution for Core42, covering GPU compute, fabric, storage, platform, and security layers.
- Designed the AI Factory GPU data-center architecture with rail-optimized InfiniBand (UFM), separate storage and in-band networks, DDN and VAST storage, and automated Cisco NDFC fabric and Palo Alto firewall provisioning.
- Authored the target architecture for Token Factory, a multi-region sovereign per-token AI inference marketplace, including metering, OpenAPI surface, Kubernetes contracts, security architecture, and conformance tests.
- Defined the build team's engineering roles for Token Factory.
- Designed KV-cache tiering and sharing with LMCache, Mooncake, and SGLang HiCache to offload KV cache from GPU memory to CPU and storage, increasing throughput and reducing time-to-first-token.
- Led technical solutioning for RFP responses to the UAE government and enterprise tenders, including architecture, bill of materials, compliance matrices, and sizing.
- Deployed a multi-modal open-weight model catalog to production featuring DeepSeek V4, GPT-OSS 120B, Qwen3-Omni, Gemma 4, and LTX-2.5 served through an APISIX gateway with per-token pricing.
- Optimized inference with NVIDIA Dynamo using disaggregated prefill/decode and KV-aware routing. Built air-gapped bundles for vLLM, the AI gateway, and AMD GPU operator with pre-flight image checks.
- Built Packer/MAAS golden images with CIS hardening for NVIDIA HGX and AMD MI210 fleets, and developed the CMP IAM layer in Go, integrating Keycloak SSO, OpenFGA, and Teleport access.
Technologies: Kubernetes, NVIDIA CUDA, AMD, InfiniBand, OpenAPI, vLLM, Packer, MAAS, Go, Keycloak, OpenFGA, Teleport, Loki, NVIDIA GPU Operator, SGLang, Helm, LXC, GPU Computing, Ceph, Grafana, Prometheus, RHEL, OpenShift, AI Solutions, Engineering, kvcache, DynamoAI, Cloud Architecture, Rados, Red Hat OpenShift Container Platform 4.10, Red Hat OpenStack Platform 13, Ansible 2.3, AI Inference, Key-Value (KV) Cache, Optimization, AI Infrastructure, GPU Clusters, Linux, NVIDIA vGPU, ROCm
Cloud Architect
2022 - 2025
Coredge.io
- Delivered private cloud for Airtel and NIC (Government of India) by designing, deploying, and operating OpenStack and Ceph.
- Built an end-to-end automation pipeline deploying a bare-metal operating system, Ceph, and OpenStack using MAAS, Jenkins, Ansible, OSA, and ceph-ansible.
- Implemented self-healing operations where Grafana alerts trigger Jenkins auto-remediation through webhooks. Automated the Prometheus/Grafana stack with Ansible.
- Tuned Ceph and OpenStack for performance and high availability. Collaborated with product engineering on IaaS/PaaS services on OpenStack and Kubernetes.
Technologies: OpenStack, Ceph, MAAS, Jenkins, Ansible, Grafana, Prometheus, OpenShift, Loki, NVIDIA GPU Operator, Helm, LXC, GPU Computing, Kubernetes, NVIDIA CUDA, AMD, Packer, Keycloak, OpenFGA, Teleport, Wazuh, RHEL, VMware vCenter, AI Solutions, Engineering, Cloud Architecture, Rados, Red Hat OpenShift Container Platform 4.10, Red Hat OpenStack Platform 13, Ansible 2.3, AI Infrastructure, GPU Clusters, Linux
Software Engineer
2020 - 2022
Snapdeal
- Deployed and operated production OpenStack and Ceph environments, installed Kubernetes with Salt, and automated processes using Jenkins and Ansible.
- Installed and configured Wazuh (SIEM) for server hardening and security benchmarking.
- Owned the complete software-defined storage stack using CEPH and supported production loads.
- Worked on creating a pipeline for Jenkins jobs to automate the S3 bucket creation using CEPH Rados.
Technologies: OpenStack, Ceph, Kubernetes, Jenkins, Ansible, Wazuh, Loki, Helm, Amazon Web Services (AWS), LXC, Keycloak, Teleport, Grafana, Prometheus, RHEL, Engineering, Rados, Red Hat OpenStack Platform 13, Ansible 2.3, Linux
Technical Service Specialist
2019 - 2020
IBM
- Deployed and supported RHEV environments. Assisted application teams in running workloads on RHEL virtual machines.
- Managed a fully air-gapped environment, ensuring robust support for the project despite the challenges.
- Enabled the sudo user auditing for all systems. Since sudo is for privilege escalation, their logs for individual commands should also be audited.
Technologies: Red Hat Enterprise Virtualization (RHEV), RHEL, Ansible, Engineering, Ansible 2.3, Linux
Cloud Consultant
2018 - 2019
Ericsson
- Maintained the telco cloud for Telenor on Ericsson CEE (OpenStack) using ECM and ScaleIO. Supported Hitachi HDS 800 and Nexenta storage.
- Achieved success in managing Nexenta storage systems, utilized for executing data center and disaster recovery scenarios.
- Automated the OpenStack workflows for cleanup, purging the DB, and restoring the instance if it went into a stale state.
Technologies: OpenStack, ECM, ScaleIO, Amazon Web Services (AWS), VMware, Ansible, Grafana, Prometheus, RHEL, Engineering, Ansible 2.3, Linux
Associate Professional
2015 - 2018
DXC Technology (Formerly CSC)
- Managed Linux and Unix server estates for Royal Mail Group and Circle K. Provisioned VMware vCenter and vCloud Director resources.
- Automated the complete daily health reports for multiple environments over email via SMTP.
- Implemented comprehensive monitoring for hypervisors and VMware VMs, ensuring optimal performance and system management.
Technologies: VMware vCenter, vCloud Director, Amazon Web Services (AWS), VMware, Ansible, Grafana, Prometheus, RHEL, Engineering, Ansible 2.3, Linux
Project History
Sovereign AI Token Factory
https://tokenfactory.iamsaif.ai/Built at Netgroup.ai and I led the architecture, A multi-region, sovereign AI inference platform that lets enterprises and government users consume open-weight LLMs through a single API, billed per token. I was the lead architect: I authored the target architecture, defined the OpenAPI surface, metering model, Kubernetes contracts, security architecture and conformance tests, and defined the engineering roles for the build team. I deployed a multi-modal model catalogue (chat, vision/OCR, embeddings, speech, image and video) behind an APISIX AI gateway with per-token metering, and optimized inference with NVIDIA Dynamo and vLLM using disaggregated prefill/decode, KV-aware routing and KV-cache offloading to CPU and storage, raising throughput and cutting time-to-first-token. Identity and access were built in Go with Keycloak SSO, OpenFGA authorization and Teleport. The platform ships as offline bundles for fully air-gapped, in-country deployment, with VictoriaMetrics, Grafana and Loki Observability.
Education
2014 - 2017
Bachelor's Degree in Computer Science
Institution of Engineers - Delhi, India
2011 - 2014
Diploma in Computer Engineering
Jamia Millia Islamia - New Delhi, India
Certifications
SEPTEMBER 2024 - PRESENT
Red Hat Certified System Administrator in OpenShift
Red Hat
MARCH 2022 - PRESENT
Red Hat Certified System Administrator in Red Hat OpenStack
Red Hat
JUNE 2019 - PRESENT
Red Hat Certified Specialist in Ansible Automation
Red Hat
MAY 2018 - PRESENT
Red Hat Certified Engineer in Enterprise Linux
Red Hat
APRIL 2016 - PRESENT
Red Hat Certified System Administrator
Red Hat
Skills
Libraries/APIs
OpenAPI, vLLM
Tools
Keycloak, Jenkins, Ansible, Grafana, Red Hat Enterprise Virtualization (RHEV), Loki, Helm, Packer, VMware
Platforms
Kubernetes, NVIDIA CUDA, AMD, OpenStack, OpenShift, Linux, Wazuh, Amazon Web Services (AWS)
Storage
Ceph
Languages
Go
Other
MAAS, Prometheus, RHEL, NVIDIA GPU Operator, LXC, GPU Computing, Large Language Models (LLMs), AI Solutions, Engineering, kvcache, DynamoAI, Cloud Architecture, Rados, Red Hat OpenShift Container Platform 4.10, Red Hat OpenStack Platform 13, Red Hat Enterprise Linux 7, Ansible 2.3, AI Inference, GPU Clusters, AI Infrastructure, Key-Value (KV) Cache, InfiniBand, VMware vCenter, vCloud Director, SGLang, Optimization, OpenFGA, Teleport, ECM, ScaleIO, NVIDIA vGPU, ROCm
Collaboration That Works
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
1
Share your needs
Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2
Choose your talent
Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3
Start your risk-free talent trial
Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.
Top talent is in high demand.
Start hiring