Stable Base is a founder-led team of senior engineers who design, run, and secure production infrastructure — from Kubernetes platforms to GPU clusters. We embed with your team, fix what's fragile, and hand over systems that are documented, automated, and owned by your people.
GPUs in a multi-cloud fleet for training and inference
concurrent players on matchmaking and game servers
AWS accounts migrated to one unified platform
services operated via GitOps by a lean SRE team
If one of these is on your roadmap or keeping you up at night, we've solved it before, at scale.
We don't hand you a slide deck. We design, build, test, and operate — inside your repos and tools — from first commit to day-two operations.
Kubernetes platforms sized to your workloads: autoscaling, mixed node pools provisioned on demand, and everything declared in code — on any cloud, hybrid, or on-prem. Moving off a legacy setup or between clouds? We plan the cut-over and run it without downtime.
GPU clusters built for distributed training and low-latency inference: scheduling, autoscaling, training pipelines, and model serving. NVIDIA (A10–H100, B200) and AMD (MI300X–MI350X), wired with RDMA, RoCEv2, and InfiniBand for cross-node throughput.
Systems that recover on their own before anyone gets paged. SLOs tied to what users feel, alert routing across time zones, and on-call your engineers can live with.
Stop guessing. Metrics, logs, and traces correlated across every cluster, with alerts tied to SLOs instead of noise — including air-gapped and on-prem environments.
Built into the platform, not bolted on before an audit. Supply-chain security, secrets management, least-privilege access, policy-as-code, and compliance controls that run continuously.
Find the holes before someone else does. Hands-on testing of web apps, APIs, cloud accounts, and Kubernetes — by engineers who build these systems, and can help you fix what we find.
All of your infrastructure in code, with drift detected and corrected automatically. Git-driven delivery where every change is reviewed, auditable, and one revert away from undone — plus self-service provisioning, so developers stop waiting on tickets.
Selected engagements and hands-on work across six industries, at the scale where shortcuts stop working.
Tens of thousands of GPUs
Architected a distributed multi-cloud GPU and CPU fleet across hyperscalers and emerging providers from scratch, supporting parallel multi-model training and inference at tens-of-thousands-of-GPU scale. Continuous deployment for API and inference workloads, distributed deployments with deep observability.
500K concurrent players
Rebuilt automation pipelines and migrated 40+ AWS accounts to a unified platform. Scaled matchmaking and game servers to 500K concurrent players with autoscaling and global edge nodes. 6,000+ services across multiple accounts, games, and environments — operated by a lean SRE team via GitOps.
Acquired by HERE Technologies
Cloud-native infrastructure for connected vehicle platforms — real-time telemetry, OTA updates, large-scale fleet management. Migrated services from on-prem with BMW and Daimler teams in Stuttgart and Berlin. Culminated in the company's acquisition by HERE Technologies GmbH.
High-performance TCP on Kubernetes
Rebuilt and migrated AR and API infrastructure to Kubernetes on AWS with custom networking and high-performance TCP services. Comprehensive documentation, on-site training in Tokyo.
Air-gapped SCADA observability
Built an observability platform for air-gapped on-prem SCADA infrastructure. Collaborated with IAM and engineering teams, supported hiring, worked on-site in Brussels.
Zero-downtime migration
Migrated high-traffic services from ECS to Kubernetes with zero downtime and faster deploys. Per-team multi-tenancy and self-hosted identity let teams ship independently and securely.
Facing something similar? Tell us where it hurts, and we'll walk you through how we'd approach it.
Book a call with an engineerStart with a scoped assessment or pen test, or bring us in to build. Every engagement begins with a written proposal: scope, timeline, deliverables.
A focused review of your architecture, reliability, observability, and security. You get a written report and a prioritized roadmap — useful whether we do the work or your team does.
A scoped, hands-on test of your applications, APIs, cloud accounts, and Kubernetes clusters, with a clear report your engineers can act on.
Our engineers join your team to design, build, and ship — in your repos, your tools, your workflows — and leave it documented and owned by your people.
Continued support and evolution after handoff — a senior pair of hands for reviews, upgrades, and the moments that matter.
A structured engagement that moves fast. We adapt to your pace, your tools, and your priorities — not the other way around.
We learn your stack, your constraints, and what's hurting — before proposing anything.
We review what you have, find the gaps, and send a written proposal with scope, timeline, and deliverables.
We join your team and build, pair, review, and ship — in your repos, your tools, your workflows.
Documentation, runbooks, and knowledge-transfer sessions, plus an optional retainer for ongoing support.
Modular infrastructure as code. GitOps delivery. Observability tied to SLOs. Security and compliance controls built into the platform.
Infrastructure your team understands and owns. Deploys in minutes, not hours. Fewer pages and faster recovery. Audit evidence ready when the auditor asks.
Stable Base was founded by Alaa, a reliability and infrastructure engineer who led SRE at Luma AI — building and running the clusters behind its training and inference systems as it scaled from a single node to tens of thousands of GPUs. Before that: microservices platforms, connected-vehicle infrastructure, and game backends serving hundreds of thousands of players.
You work directly with hands-on expert engineers — people who design, build, and operate production systems every day, and who stay close to the work from first call to handoff.
Alaa on LinkedInColleagues from AI infrastructure at scale to microservices platforms. Read the full recommendations on LinkedIn.
I've worked alongside Alaa at Luma AI, where he leads our SRE function. He built and maintained the core infrastructure that powers both our training and inference systems — work that is foundational to everything we do as a company.
"Alaa is a true powerhouse SRE. In the early days of Luma he was solely responsible for managing our clusters, and he worked tirelessly to keep everything reliable while we scaled. He maintained a very high bar for both our infrastructure and the calibre of new hires."
"Worked with Alaa at Luma, where he headed the SRE organization. From managing massive GPU clusters to diagnosing issues at scale, Alaa was instrumental in scaling AI infrastructure at the company from the early days. I learned a lot about GitOps, observability and debugging subtle issues from my time working with him."
"Through his leadership of the SRE team we have been able to accomplish great things. He has a deep focus on making sure our infrastructure is secure and fully automated. A phenomenal reliability engineer that can lead and architect top of the line systems."
"One of the hardest working SRE / AI infrastructure folks at Luma. He helped scale our resources from a single node to thousands of nodes across multiple backbones. Alaa has been a crucial part of Luma's success allowing us to effectively scale our resources and compute."
"One of the best engineers I've enjoyed working with. He built whole infra in Luma from scratch, made some impossible things possible."
"Alaa set up, maintained, and built tools for our GPU infrastructure on multiple cloud providers across tens of thousands of GPUs in a maintainable and reliable way. Alaa also has very strong cross-functional intuition and goes above and beyond to build systems to the needs of internal teams and external customers alike."
"Alaa established most of our infrastructure on Kubernetes. He worked closely with developers and made it easy to deploy and scale services up and down. He also implemented an observability stack on all services."
"Extremely skillful and passionate engineer that enjoys building scalable and reliable infrastructure/solutions. Willing to share his knowledge and mentor others. It was a pleasure working with him."
"I've worked closely with Alaa for a few months on an online multiplayer gaming backend project. His passion for his craft is contagious, and he is never shy to share his knowledge and expertise."
"Within a few weeks he managed to build a scalable yet reliable framework for the infrastructure team to build on and effectively reduce operational costs from weeks to hours. He keeps documentation up to date for others to follow."
"His wide knowledge across the whole stack — together with a deep understanding of distributed systems, algorithms and protocols underlying the applications we worked on together — makes him a truly versatile problem-solver."
"Alaa managed our company's infrastructure on AWS. He is really good at using abstraction and automation to scale platforms and environments to any size. A quick and eager learner always implementing the latest best practices."
"In almost two years working with Alaa I can't remember a single problem with the infrastructure he has built and maintained. Constantly striving for improvement of the infrastructure."
"His continuous delivery of highly reliable infrastructure in a fast-moving environment was a key part of the success of our startup. A fount of knowledge and an all-round great person to work with."
"At Brainly we created, led by Alaa, an immutable, scalable and highly tolerant internal microservices platform able to run thousands of docker based units. If you need a modern but still strong and reliable platform, Alaa is one of the best bets I know."
"He took an extra mile to introduce and implement his ideas on how we could improve the infrastructure at Wimdu. A very nice person, easy to work with, who can quickly integrate with the team."
"Alaa knows unix-based systems inside out. During his time at Wimdu he managed to improve existing infrastructure a lot and has been the main innovation driver in this area. The kind of challenges that would be overwhelming are welcomed by him with excitement."
"Alaa is very proficient and has deep understanding of computer system security what enables him to do magic. He grasps new tools and technologies with ease."
"A truly exceptional devops engineer, always at the forefront of technology, not afraid to push the boundaries, with stability and security always at the forefront of his thinking. At Brainly he developed a highly scalable and highly redundant immutable infrastructure on which we built microservices."
"He successfully implemented a microservices platform. It is extremely resilient to failure and self-healing. If you ask me if I want to work together with him, my answer is: 'Yes, anytime.'"
End-to-end post-quantum encrypted secret sharing for teams.
Built for the operational data teams still pass around in Slack, email, and tickets — credentials, kubeconfigs, tokens, logs, and files. Every secret is encrypted on the sender's device for specific recipients. Private keys never leave the device; the server only ever stores ciphertext.
Something else on your mind? Email us and an engineer will answer.
With it. We embed in your repos, tools, and workflows, pair with your engineers, and finish with documentation and knowledge-transfer sessions so your team owns what we build.
The major hyperscalers, emerging GPU providers, and on-prem — including air-gapped environments. Kubernetes, infrastructure as code, and GitOps sit at the core of most of our work.
Yes. We build security and compliance into the platform itself — supply-chain security, secrets management, policy-as-code, and automated compliance controls for SOC 2, ISO 27001, ISO 42001, and GDPR.
We agree the scope up front — web apps, APIs, cloud accounts, Kubernetes. You get a report with severity-ranked findings, reproduction steps, and concrete fixes, and we can help your team remediate.
Every engagement is quoted in a written proposal before any work starts, so scope, timeline, and cost are agreed up front. Assessments and pen tests are scoped to a fixed set of deliverables.
With a discovery call — book one here. Then an assessment and a written proposal with scope, timeline, and deliverables, before any work begins.
Sheridan, Wyoming. We work with teams globally, Monday to Friday, and have worked on-site in Europe and Asia when the job called for it.