About
Articles by Walid
-
🚀 OpenShift & AWS: Maximizing Cost Efficiency
🚀 OpenShift & AWS: Maximizing Cost Efficiency
For those running ROSA "Red Hat Openshift Service on AWS" with Hosted Control Planes (HCP), 4.21 introduces critical…
Activity
15K followers
-
Walid Shaari reposted thisWalid Shaari reposted thisBeen working on a fun weekend project. My brother runs a large crew of pipelayers. The math they do is genuinely hard, and I don't understand any of it. But I know how to build apps, so I built him a calculator PWA. Problem is, every change request went through me. "Make the font bigger." "The offset calculator is wrong for 22.5 degree fittings." I'd context-switch, misunderstand what he meant, and push a fix days later. I was the bottleneck for an app where I didn't even know the domain. So I built a Telegram bot that lets him describe what he wants in plain English. An AI agent makes the changes, validates them, deploys a preview, and waits for his approval before pushing to production. He never touches code, git, or a terminal. Wrote up the full architecture and the decisions behind it: https://lnkd.in/ebJir3sC #AIAgents #AWS #Serverless #BedrockAgentCoreMy Brother Doesn't Code. Now He Ships Features. — edjgeekMy Brother Doesn't Code. Now He Ships Features. — edjgeek
-
Walid Shaari reposted thisWalid Shaari reposted thisI’m #hiring. Looking for an Application Modernization Architect to join our team in London! Know anyone who might be interested? Please share this post with your network!
-
Walid Shaari reposted thisWalid Shaari reposted thisالاخوة السوريين لو محتاجين تتابعوا شباب سوري على LinkedIn بيقدموا محتوى و منشورات مهمه في مجال ال Tech , انصحكم بـــ : Ghadeer AlHasan محتوى خاص بالبرمجيات و المسارات المهنية Mahmoud Khatib محتوى للكلاود Philip Nakhleh محتوى ذكاء اصطناعي Rahaf Al-Najjar محتوى ذكاء اصطناعي وكلاود Wael Hasan محتوى Microsoft Azure يا رب اكون دبستهم علشان ينشطوا اكتر 😁
-
Walid Shaari reposted thisWalid Shaari reposted thisLast week I was chatting with my AWS colleague Brian Beach and complimented him on his string of LinkedIn posts that documented and explained various features of Kiro. Yesterday I decided to use Kiro to find those posts and to create a simple HTML document pointing to them. Read my new post, "Fun with Kiro: LinkedIn Scanner" to see how I did it: https://lnkd.in/gRRebDmN
-
Walid Shaari reposted thisGreat opportunityWalid Shaari reposted this🚀 PhD & Postdoctoral Opportunities in Agentic AI for Networking and Cloud Computing at KFUPM, Saudi Arabia I am pleased to announce multiple PhD and Postdoctoral positions at the King Fahd University of Petroleum and Minerals (KFUPM), Saudi Arabia. These positions offer the opportunity to work at the Interdisciplinary Research Center for Intelligent Secure Systems (ISS) at KFUPM on cutting-edge research within the FlexNGIA 2.0 project focusing on cutting-edge research on Agentic AI for network and service management, cloud computing, and next-generation Internet architectures. To learn more about the project, please check the FlexNGIA 2.0 paper (https://lnkd.in/gYjbQbnj) or visit the project website (www.FlexNGIA.net). Benefits: - PhD positions: Fully funded for 4 years, providing strong support to pursue your research in a dynamic and innovative environment. - Postdoctoral positions: Tax-free competitive salary, health insurance, and access to excellent research facilities and collaboration opportunities. Eligibility and Required documents: - PhD applicants: CV, Academic transcripts, Statement of purpose (related to FlexGIA 2.0), 3 recommendation letters - Postdoc applicants: PhD awarded in 2023 or later, CV, Research statement, List of publications, 3 recommendation letters 📩 Application Submission: Please send your complete dossier to Prof. Mohamed Faten Zhani (mohamed.zhani@kfupm.edu.sa) #PhD #Postdoc #AI #Networks #CloudComputing #AgenticAI #Research #KFUPM #SaudiArabia
-
Walid Shaari reposted thisWalid Shaari reposted thisAmazon S3 Files just launched, and it changes how Lambda functions work with data. You can mount an S3 bucket as a local file system. Your Lambda code uses open() and pathlib instead of boto3. I built a working example: AI code review agents that share a mounted S3 workspace. A durable function orchestrator clones a GitHub repo, then two Strands agents (security + style) review the code in parallel. All through standard file I/O on the mount. The repo includes a full SAM template with everything you need to deploy this yourself. S3 Files CloudFormation is brand new and underdocumented, so I wrote up every IaC gotcha I hit: the trust policy, access point CreationPermissions, async API Gateway for durable functions, the works. Four AWS features in one deployable example: → S3 Files (mount S3 as a file system) → Lambda (serverless compute) → Durable Functions (orchestration with checkpointing) → Strands Agents SDK (AI agents with Bedrock) https://lnkd.in/erUyek6k #serverless #aws #lambda #s3files #aiagents #iacLambda Just Got a File System. I Put AI Agents on It. — edjgeekLambda Just Got a File System. I Put AI Agents on It. — edjgeek
-
Walid Shaari reposted thisWalid Shaari reposted this🚀 Excited to announce Amazon ECS Managed Daemons for ECS Managed Instances! ECS customers can now deploy and manage software agents — security, observability, networking — across their container infrastructure, independently from application deployments. Managed daemons enables platform teams to maintain consistent agent coverage and lifecycle management independently from application operations. Key benefits: ✅ Guaranteed coverage — exactly one daemon task on every instance, starting before application tasks and draining last, so your agents are always running when your applications need them ✅ Independent updates — update daemons on their own schedule with rolling deployments, circuit breaker protection, and automatic rollback — no application redeployments needed ✅ Efficient resource use — one shared daemon per instance replaces a sidecar model of running a daemon alongside every application task ✅ Auto-repair — if a daemon becomes unhealthy, ECS automatically drains and replaces the instance Available now in all AWS commercial regions at no additional cost—you pay only for the compute resources your daemon containers consume. To learn more, see: 📰 What's New post: https://lnkd.in/gdQERGCn 📝 Blog post: https://lnkd.in/g94_hmZ5 📚 Documentation: https://lnkd.in/gEm7HRjU Would love to hear your feedback as you try it out! 🎯 #AWS #ECS #Containers #CloudNative #PlatformEngineering
-
Walid Shaari reposted thisWalid Shaari reposted thisClaude Mythos is Anthropic's new model built specifically for finding security vulnerabilities in code. It just landed on Amazon Bedrock! 🛬 It's part of Project Glasswing, but don't get too excited yet. It's gated behind an allow-list, currently only in us-east-1, and AWS account teams are contacting eligible orgs directly 😢 Mythos is meant to reason about actual exploitability, reading large codebases, identifying real vulnerabilities, and understanding attack paths with less manual guidance than previous models. It's a completely separate model class from Sonnet/Opus. Anthropic calls it "fundamentally new" (whatever that means 😬) As I haven't gotten my hands on it yet, so I can't tell you how it performs in practice 🤷♂️ What I do know is this is the first time a model has been _purpose-built_ for offensive/defensive security reasoning at this level and shipped directly into a cloud provider's model catalog 👀 Worth watching even if you can't access it today. Link to the announcement: https://lnkd.in/dYqrG-WF Link to Anthropic's announcement: https://lnkd.in/dAUWh2cm P.S. Security on AWS starts with IAM. If you want a visual breakdown of how it all fits together, our IAM infographic covers the core concepts: https://lnkd.in/dn-V68yv
-
Walid Shaari reposted thisWalid Shaari reposted thisIf you work with ML models, Python packages, npm modules, or Rust crates, your machine is quietly hoarding gigabytes of cached data you forgot about. Some of it is outdated. Some of it has known CVEs. Most of it is invisible. Happy to share my new project, Cache Commander: a terminal tool for exploring, auditing, and cleaning developer caches on macOS and Linux. ✅ Hugging Face models and datasets (hub, xet): often tens of GBs ✅ uv / pip wheels and sdists: old versions with unpatched vulnerabilities ✅ npm / npx: transitive dependencies with install scripts hiding in cache ✅ Cargo registry and crate archives ✅ Homebrew bottles and downloads ✅ PyTorch hub, ChromaDB embeddings, etc. ✅ Easily extensible to additional caches (what are you missing?) For each cached package, ccmd can: ✅ Scan for known CVEs via OSV.dev ✅ Check for newer versions from package registries ✅ Show deletion safety ccmd is built with Rust, so it's tiny and fast. You can scan 50+ GB of caches in seconds, not minutes. Install: brew install ccmd - cargo install ccmd - cargo binstall ccmd GitHub: https://lnkd.in/eX-kMQ66 Please give it a ⭐️ if you like it!
-
Walid Shaari liked thisWalid Shaari liked thisToday I'm happy to ship something I've been building for a while: Canopy, a native macOS app for running hyper-productive parallel Claude Code sessions with git worktrees. Here's the problem. Claude Code is a superpower. But real engineering work doesn't arrive one task at a time. A bug lands while you're refactoring. A review request shows up while you're writing tests. You need to work on three things at once, but Claude and Git want you to work on one. Sure, you can hack around it. Stash, checkout, a second clone, maybe `git worktree add`, a shell script to copy your `.env`. What you get is a Terminal tab graveyard, stale worktrees rotting on disk, Claude sessions you can't find. A quiet tax on every task. Canopy removes that tax. ➡️ One window. Parallel worktree sessions in tabs. ⌘1–9 to jump. ➡️ Claude conversations auto-resume across restarts. No --resume flags, no session IDs to hunt down. ➡️ Per-project config: files to copy, directories to symlink, setup commands to run on every new worktree. ➡️ One-click Merge & Finish. No more forgotten commits or stale worktrees rotting on your laptop. ➡️ Activity dashboard: token usage, model split, 12-week heatmap of your actual Claude hours. ➡️ Split terminal, find-in-terminal, IDE integration, and more: the things that save you 30 seconds every time you use them. Native SwiftUI. No Electron. Signed, notarized, free under AGPL-3.0. ✅ Download (.dmg): https://lnkd.in/eQvgHHQp ✅ Homebrew: brew install --cask juliensimon/canopy/canopy ✅ Source & docs on Github - link in the comments If you use Claude Code seriously, please give it a try, drop a ⭐, and tell me what to build next.
-
Walid Shaari liked thisWalid Shaari liked thisGPU autoscaling for LLM workloads on Kubernetes is a real gap in most setups. Standard CPU and memory metrics tell you almost nothing about whether your model is actually overwhelmed, which means teams end up either over-provisioning expensive GPUs or discovering problems too late. This article walks through the full pipeline including deploying Prometheus on GKE Autopilot, wiring up vLLM metrics via a cross-namespace ServiceMonitor, fixing a Grafana dashboard to match actual metric names, and building an HPA that scales on request queue depth instead of CPU. Akrem Issaoui wrote this up with gotchas and workaround included, from Autopilot's host-level restrictions to the colon vs underscore metric naming trap. Check it out if you are running GPU workloads on Kubernetes and want autoscaling that responds to what actually matters. https://lnkd.in/ein6ZJZjGPU Autoscaling on Kubernetes: From Prometheus Metrics to HPA with vLLM on GKEGPU Autoscaling on Kubernetes: From Prometheus Metrics to HPA with vLLM on GKE
-
Walid Shaari liked thisWalid Shaari liked thisComputer now connects with Plaid to link bank accounts, credit cards, and loans. Track spending in detail, build custom budget tools, and visualize your net worth alongside your investment portfolio. All users can link their accounts, access the portfolio page, and ask Perplexity about their finances. Computer tasks are exclusive to Pro and Max subscribers. Read more: https://lnkd.in/gHvs-kmU
-
Walid Shaari liked thisWalid Shaari liked thisI've been using my Kiro Kantoku tool to manage my Kiro 👻 work more and more, and I'm now using it to make changes to the tool itself.Having a UI over all the CLI agent sessions has helped massively, but what if you could go a step further?! I've just added the ability to view all your Kiro Agents working inside your very own Kiro Haunted Office! Agents that are working are at their desk 💻 . Agents that are waiting for more work are chilling with some coffee ☕ Thought this was a fun way to visualise your Kiro team getting on with some work. The House of Kiroween never closes since I've been using Kantoku to manage the agents, I've been able to manage tasks that roll over a day or so, that I forget about, and can easily come back to. Combined with the new 1 million token context window on Sonnet and Opus, it's a great match! https://lnkd.in/eFn-AbP4 #kiro #claude #aws #agenticcoding
-
Walid Shaari liked thisWalid Shaari liked thisObservability is very important! Deploying containers to Kubernetes is the easy part. Understanding how those applications perform, how resources are used, and whether the system is healthy - that's the real challenge. The answer is building observability into the platform from day one, not bolting it on later. This example below demonstrates an observability-first EKS platform using modular Terraform. Prometheus and Grafana are deployed before workloads, not after. The focus is Day-2 operations: metrics, CPU-based autoscaling with HPA, and failure recovery. Tanushree Aggarwal includes the full implementation with Terraform modules for VPC, EKS, observability, and apps. A GitHub repo included. Check it out! https://lnkd.in/eXmXY6VFObservability-Driven Kubernetes: A Practical EKS DemoObservability-Driven Kubernetes: A Practical EKS Demo
-
Walid Shaari liked thisWalid Shaari liked this🚀 Most platforms break at scale, not because of bad tools, but because of copy-pasted pipelines, drift, and zero governance. In this video, I break down the 3 core platform engineering challenges and how Octopus Deploy's Platform Hub solves them. Full video below 👇🏼
-
Walid Shaari liked thisWalid Shaari liked thisHuge milestone from Red Hat today. OpenShift Sandboxed Containers 1.12 and Red Hat build of Trustee 1.1 just graduated from Tech Preview to General Availability to enable confidential computing on bare metal. This is a product our team helped co-engineer with some of our largest customers. Confidential computing is the basis for multi-tenancy, and a requirement for sovereign infrastructure. It means hardware-based memory encryption and attestation, not just in the cloud, but on-prem too. And now there's Technology Preview support for NVIDIA Confidential Computing, bringing that same isolation to GPU-accelerated AI workloads. This is the kind of infrastructure trust layer enterprises need as they run sensitive AI workloads at scale. https://lnkd.in/eBWjib-9 #RedHat #OpenShift #ConfidentialComputing #AI #HybridCloud #EnterpriseAI #OpenSourceRed Hat OpenShift sandboxed containers 1.12 and Red Hat build of Trustee 1.1 bring confidential computing to bare metal and AI workloadsRed Hat OpenShift sandboxed containers 1.12 and Red Hat build of Trustee 1.1 bring confidential computing to bare metal and AI workloads
-
Walid Shaari liked thisWalid Shaari liked thisBeen working on a fun weekend project. My brother runs a large crew of pipelayers. The math they do is genuinely hard, and I don't understand any of it. But I know how to build apps, so I built him a calculator PWA. Problem is, every change request went through me. "Make the font bigger." "The offset calculator is wrong for 22.5 degree fittings." I'd context-switch, misunderstand what he meant, and push a fix days later. I was the bottleneck for an app where I didn't even know the domain. So I built a Telegram bot that lets him describe what he wants in plain English. An AI agent makes the changes, validates them, deploys a preview, and waits for his approval before pushing to production. He never touches code, git, or a terminal. Wrote up the full architecture and the decisions behind it: https://lnkd.in/ebJir3sC #AIAgents #AWS #Serverless #BedrockAgentCoreMy Brother Doesn't Code. Now He Ships Features. — edjgeekMy Brother Doesn't Code. Now He Ships Features. — edjgeek
Experience & Education
-
Amazon Web Services (AWS)
View Walid’s full experience
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Licenses & Certifications
-
ISSAF - Penetration Testing Qualified I-PTQ
ISSAF
Issued -
-
-
-
-
Offensive security wireless professional OSWP
Offensive security
Issued Expires -
Red Hat Certified Specialist in Linux Performance Tuning
Red Hat
Issued ExpiresCredential ID 805008798633161 -
-
1 STAR - SANS GIAC Skills test, and report for “Cutting Edge Hacking Techniques”
SANS Institute
Issued Expires -
Publications
-
CONTAINERS WORKSHOP DURING SAUDI HPC 2016 : DOCKER 101
See publicationworkshop slides from Saudi HPC 2016 http://www.hpcsaudi.com/ covering the Theory and hands-on of Docker containers, and Its eco system. The foundations of the Docker platform, including an overview of the platform system components, images, containers and repositories, installation , using Docker containers from repositories e.g. dockerhub, how to create a container using Dockerfile, containers development life cycle. The strategy is to demonstrate through "live demo, and shared exercise" the…
workshop slides from Saudi HPC 2016 http://www.hpcsaudi.com/ covering the Theory and hands-on of Docker containers, and Its eco system. The foundations of the Docker platform, including an overview of the platform system components, images, containers and repositories, installation , using Docker containers from repositories e.g. dockerhub, how to create a container using Dockerfile, containers development life cycle. The strategy is to demonstrate through "live demo, and shared exercise" the reuse and customization of components to build a distributed system case service gradually
Projects
-
ECC 2012 Migration to Univa Grid Engine HPC workload management and scheduler
Linux team technical lead in migrating several HPC cluster silos that were run by TORQUE/MAUI to UNIVA Grid Engine. after migration almost all cluster resources are controlled by three UGE masters in global three clusters instead of the 10s of silos. achieved better reliability, more flexibility, and better accounting. worked very closely with vendors LSF, UNIVA, and Clustervision first during the contracting and open bidding phase, and with application support department, and system for the…
Linux team technical lead in migrating several HPC cluster silos that were run by TORQUE/MAUI to UNIVA Grid Engine. after migration almost all cluster resources are controlled by three UGE masters in global three clusters instead of the 10s of silos. achieved better reliability, more flexibility, and better accounting. worked very closely with vendors LSF, UNIVA, and Clustervision first during the contracting and open bidding phase, and with application support department, and system for the final deployment under the Project Lead Hussain Raqa.
Other creators -
-
ECC Oracle RAC 2010
Linux systems representative in running factory acceptance testing, as well deploying a maximum availability Oracle RAC Linux clusters between three near-by physical sites to achieve business continuity , high availability and disaster recovery.
Other creators -
-
ECC 2007 512 Infiniband Linux node cluster
Deploy the first 512 nodes infiniband cluster for ECC to serve petroleum users reservoir simulation jobs. scalability of specific mpi, storage performance with recently released RHEL5 among other technical and people challenges were faced.
-
Linux High Performance Compute Clustering
Integration and Deployment of 19 Linux Cluster in Saudi ARAMCO Exploration Department.
4 Linux HPC Cluster with 512 computes nodes (Dell PE M610 & PE 1950 servers); using InfiniBand High Performance Compute Cluster with data transfer rate of up to 40 GBPS. 2 Qlogic 12800 switch, 2 CISCO SFS InfiniBand switches & SilverStorm 9240 InfiniBand switches are utilized. Cisco Nexus 5000 and Dell PowerConnect Switches are used in these clusters
These clusters are managed through…Integration and Deployment of 19 Linux Cluster in Saudi ARAMCO Exploration Department.
4 Linux HPC Cluster with 512 computes nodes (Dell PE M610 & PE 1950 servers); using InfiniBand High Performance Compute Cluster with data transfer rate of up to 40 GBPS. 2 Qlogic 12800 switch, 2 CISCO SFS InfiniBand switches & SilverStorm 9240 InfiniBand switches are utilized. Cisco Nexus 5000 and Dell PowerConnect Switches are used in these clusters
These clusters are managed through Platform/Scali Manage & Cluster Vision application to deploy/monitor & manage Clusters.
1 Linux HPC Cluster with 128 computes nodes (Dell PE M610 & PE1950 servers) using SilverStorm 9120 InfiniBand Switch.
6 Linux HPC Cluster with 128 compute nodes (Dell PE1950 server) for Seismic Application
3 Simulation Cluster using NVidia External GPUs for Graphics Simulation of Seismic activity.Other creators -
Languages
-
English
Native or bilingual proficiency
-
Mandarin
Elementary proficiency
-
Arabic
Native or bilingual proficiency
Organizations
-
SANS
GIAC Advisory Board
- PresentThe GIAC Advisory Board is a private group that shares information security best practices and strategies. We also advise GIAC and SANS on new topics for certification and give feedback on existing certification programs. Membership is restricted to those who have scored 90% or higher on a GIAC information security exam.
Recommendations received
18 people have recommended Walid
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content