Available for new opportunities

ACHRAFBELARIF

AI Infrastructure / Linux Engineer

>AI Infrastructure Engineering

@Oracle Labs|Casablanca, Morocco

✉  achraf.belarif@aol.com⌥  GitHub⌘  LinkedIn
Scroll
01.

ABOUT_ME

I am an AI Infrastructure Engineer specializing in the reliability, qualification, and recovery lifecycle of bare-metal GPU/HPC clusters powering large-scale AI/LLM training capacity in a hyperscale data center, with hands-on Linux virtualization experience (QEMU/KVM/Libvirt) applied directly to qualifying AI infrastructure platforms ahead of rollout, and backend engineering experience building web services that integrate with that same hardware and systems infrastructure.

With a foundation in C/C++, Python, and Bash, I specialize in understanding complex system interactions across the AI infrastructure stack, from kernel internals and hypervisors down to RDMA/InfiniBand fabrics, NVLink topologies, and physical hardware, isolating genuine faults from transient automation issues and turning validation evidence into repair-ready technical records for AI data center capacity.

Actively seeking new opportunities

AI Infrastructure & GPU/HPC Reliability

—Hardware Qualification & Triage: Validating bare-metal GPU/HPC hosts across an internal hardware validation test suite, isolating genuine hardware faults from transient automation/control-plane issues
—RDMA & Fabric Diagnostics: InfiniBand/RoCE fault isolation, mapping logical failures to NIC, PCIe BDF, cable/transceiver, and T0 switch-port paths
—GPU & Fabric Health: NVLink/NVSwitch topology, Xid events, NVIDIA Fabric Manager, PCIe/AER kernel diagnostics
—Hardware Recovery: Deep power cycle, reprovisioning, cable/NIC/GPU replacement, and RMA coordination with repair-ready evidence records
—AI Data Center Design: Rack-scale GPU topology, liquid cooling fundamentals, firmware/new-capacity enablement, fleet lifecycle management

Backend & Systems Engineering

—Backend Engineering: FastAPI, NestJS, Dropwizard, Oracle JET, and web services integrated directly with hardware and AI infrastructure systems
—Databases: PostgreSQL, Oracle Database, SQL, Prisma
—Systems Programming: C/C++, Java, JavaScript, low-level APIs, memory management, concurrency
—Scripting & Automation: Python, Bash for AI infrastructure tasks and tooling
—Container Technologies: Docker, Podman, container orchestration with Kubernetes
—Infrastructure Tools: Git, Jira, CI/CD pipelines, monitoring systems

AI Infrastructure Virtualization

—System Performance: Process scheduling, memory management, I/O subsystem tuning
—Fabric & GPU Diagnostics: rdma link, mlxlink, ethtool, ibv_devinfo, lspci, nvidia-smi for AI fabric fault isolation
—Firmware & Provisioning: PLDM firmware updates, NVMe provisioning, host discovery and lifecycle management for AI capacity
—Virtualization Interfaces: KVM/QEMU, Libvirt, live migration — validated ahead of AI infrastructure GPU platform rollouts
—Secure & Confidential Computing: Guest isolation, attestation workflows, confidential computing concepts for multi-tenant AI infrastructure
—Automated Testing: Python/Bash/Avocado frameworks for AI infrastructure regression and release-readiness validation

Professional Focus

—AI Data Center Reliability: Qualification and recovery lifecycle for hyperscale GPU capacity powering large-scale AI/LLM training
—Systems Engineering: Designing robust, scalable AI infrastructure solutions
—AI Platform Validation: Automated testing frameworks for AI infrastructure platforms, reproducible environments
—Root-Cause Analysis: Debugging complex system failures across stack layers, from software down to physical fabric
—Backend Web Development: Extended the Oracle Contributor Agreement (OCA) Signing Service (Dropwizard/Oracle JET) with new features, performance improvements, and bug fixes for external open-source contributor onboarding
—Tool Development: Custom utilities for system monitoring, testing, and automation
—Documentation & Knowledge Sharing: Technical writing, mentoring, best practices

Projects & Initiatives

—V12n-AI-Lab — AI Automation for Virtualization: Agentic loop with custom tools and skills under a multi-agent architecture, automating end-to-end AI infrastructure virtualization validation workflows
—Linux Lab Infrastructure: Personal test environments for kernel modules, container networking
—Automation Frameworks: Custom tools for environment setup, log analysis, metric collection
—Systems Research: Deep dives into kernel features, performance optimization techniques
—Open Source Contributions: Patches, documentation, community engagement

Professional Attributes

—Systems Thinking: Holistic approach to complex technical problems
—Deterministic Reliability: Focus on reproducibility, predictability in infrastructure
—Rigorous Documentation: Clear, actionable technical communication
—Trilingual Communication: Native Arabic, professional French/English fluency

Skills Matrix

AI Infrastructure · AI Infrastructure Virtualization · Backend Engineering

HPC & GPU Infrastructure

NVIDIA Blackwell (B200/GB200/GB300)AMD Instinct (MI300X/MI355X)RDMAInfiniBand/RoCENVLinkNVSwitchPCIe

AI Data Center Design

AI/HPC Data Center ArchitectureRack-Scale GPU TopologyLiquid Cooling FundamentalsFleet Lifecycle Management

AI Infrastructure Virtualization

Linux/UnixOS InternalsQEMUKVMLibvirtLive MigrationFirmwareDockerPodmanKubernetes

Networking & Security

VLANRoutingFirewallVPNDNS/DHCPGuest IsolationAttestationConfidential Computing

AI & Automation

Agentic AIMulti-Agent ArchitectureAI Infrastructure Automation

Backend & Programming

CC++JavaJavaScriptTypeScriptPythonBashFastAPINestJSDropwizardOracle JETReact

Databases

PostgreSQLOracle DatabaseSQLPrisma

Testing, Diagnostics & Automation

Hardware Validation Test Suitesmlxlinkethtoolnvidia-smirdma linkibv_devinfolspciAvocadoGitJira

Languages

ArabicEnglishFrench
02.

PROJECTS

AI infrastructure automation, Linux virtualization tooling, and backend engineering. From an agentic AI framework driving virtualization validation to systems-level projects in C/C++.

01

V12n-AI-Lab

Agentic AI automation system for the end-to-end virtualization workflow. Multi-agent framework that autonomously plans and executes AI infrastructure validation and testing tasks.

02

Fancy Logger

Web dashboard for remote log and debug data collection across Linux-based AI infrastructure hosts. Centralized log streaming, search, and diagnostic snapshots; adopted by the audit service.

03

AI Job Hunter

Semi-autonomous LinkedIn job-hunting system. An agentic AI worker handles job discovery and recruiter outreach on interleaved schedules, backed by a Next.js/Express/Prisma dashboard for human review.

04

ft_transcendence

Real-time multiplayer Pong platform with live chat, matchmaking, 42 API OAuth, 2FA, and spectator mode.

05

ft_containers

C++ STL containers (vector, map, stack, set) and iterators reimplemented from scratch with full standard compliance.

06

webserv

HTTP/1.1 server in C++98 with I/O multiplexing, CGI support, virtual hosts, and chunked encoding.

07

ft_services

Kubernetes cluster with Nginx, WordPress, Grafana, MySQL, InfluxDB, and FTPS orchestrated via k8s manifests.

08

ft_server

Debian Docker container running Nginx, WordPress, phpMyAdmin, and MariaDB from bare configuration.

09

minishell

POSIX-compatible Unix shell with pipes, redirections, environment variables, heredoc, and built-ins.

View all on GitHub
03.

RESUME

Open in DriveDownload PDF