Alex Wicks

Senior Infrastructure & Systems Engineer

London, GB github.com/aw1cks alex+cv@awicks.io +44 7484 616643 cv.awicks.io

Profile

Infrastructure and systems engineer with 7+ years designing, building, and operating Linux, Kubernetes, GPU compute, and networking platforms across AI and financial services. Focused on complex cross-layer systems engineering — from kernel internals and CPU scheduling to distributed infrastructure — combining deep operational expertise with practical software development in Rust, Go, Python, and Kotlin. Experienced in platform architecture, technical leadership, and building reusable capabilities adopted across engineering teams.

Experience

Maven Securities

Platform Engineer

Jun 2024 – Present
  • Engineer and operate Kubernetes and internal developer platforms across 20+ clusters for ~100 engineers, driving architecture, upgrades, and cross-team technical reviews.
  • Diagnosed Linux kernel failures through vmcore forensics and upstream/backport analysis; built custom kernel packaging and CI, delivering improved memory reclaim through MGLRU, reduced kernel memory consumption through folios, and resolution of an eBPF memory leak affecting Cilium.
  • Architected multi-region LiteLLM gateway infrastructure for regional latency reduction and failover, using Cilium BGP anycast, health-based route withdrawal, horizontal autoscaling, and regional Redis; optimised Kubernetes CPU/NUMA policy and storage for H200-hosted vLLM workloads.
  • Extended a Go Kubernetes operator across 20+ clusters for Keycloak/AD reconciliation and Envoy Gateway OIDC; built Rust Kerberos/SPNEGO cluster auth tooling.
  • Implemented automated internal PKI with step-ca and ACME across hundreds of non-Kubernetes services, eliminating manual certificate renewals.
  • Built cross-platform Rust CI pipelines (Linux/Windows, Artifactory) in Kotlin TeamCity tooling, adopted by another team for production trading software.
  • Diagnosed VMware CPU contention and excessive vNUMA migrations using hypervisor metrics and Linux PSI; optimised VM topology to reduce contention and improve GitHub Enterprise responsiveness.
  • Migrated logging to Strimzi/Kafka on Kubernetes (cutting lag from hours to minutes) and deployed VAST S3/Trino/Hive, retiring ~$200k/year in AWS Athena spend.

Speechmatics

Senior Site Reliability Engineer

Jan 2023 – Apr 2024
  • Integrated NVIDIA H100 nodes into bare-metal AI clusters, including GPU telemetry, Sun Grid Engine source patches, and custom Debian packages for Ubuntu 22.04.
  • Reconstructed undocumented on-premises network topology and diagnosed core-switch buffer overruns via Scapy and SNMP, driving uplink and switching-capacity remediation.
  • Led a datacentre migration to a new colocation facility to unlock power and cooling capacity for GPU expansion; delivered internet, firewall, and Azure VPN connectivity in one month via a Megaport-hosted Fortinet appliance, later establishing ExpressRoute for cloud bursting and CI.
  • Delivered the first end-to-end SaaS release pipeline by assembling GitLab CI, Terraform, and Argo CD into an operational workflow, cutting release effort from days to hours.
  • Automated Ubuntu golden images via cloud images, Packer/QEMU, and Python/govc; built unattended Windows 11 images and Ansible provisioning for GitLab CI runners.

Site Reliability Engineer

Jan 2022 – Jan 2023
  • Managed a 20–30 host VMware estate running hundreds of VMs and migrated vSAN storage to a Pure SAN, resolving iSCSI, MTU, and I/O contention bottlenecks.
  • Packaged VAST CSI into a Helm chart and built reusable Terraform modules with Terratest conformance checks for self-service infrastructure.
  • Built network and platform observability tooling (SNMP, LLDP, blackbox monitoring) supporting internal AI research and training platforms for ~50 engineers and ML researchers.

Capgemini — HMRC Market Unit

Platform Engineer, Metrics & Monitoring

Jul 2021 – Dec 2021
  • Engineered centralised Prometheus and Thanos observability across thousands of AWS EC2 instances and Kubernetes workloads for long-term metric retention.
  • Implemented blackbox monitoring for critical Consul and Vault services and migrated monitoring components from legacy EC2 hosts onto Kubernetes.

Flextrade Systems

DevOps Engineer

Aug 2020 – Jul 2021
  • Built bare-metal Kubernetes clusters via PXE automation and migrated CI workloads onto Kubernetes, cutting build times by ~40%.
  • Designed Jenkins, Python, and Ansible delivery automation replacing manual releases, and supported VMware Cloud on AWS hybrid infrastructure.

Systems Administrator

Jan 2020 – Aug 2020
  • Automated VM provisioning with Python tooling, managed vSphere lifecycle upgrades, and mentored junior engineers.

Junior Systems Administrator

Mar 2019 – Jan 2020
  • Supported multi-region production SaaS infrastructure across Linux, Windows, VMware, storage, networking, Active Directory, databases, and firewalls, including on-call incident response.

Selected Projects & Open-Source

Linux LD_PRELOAD shim aligning glibc CPU-discovery interfaces (sched_getaffinity, sysconf, get_nprocs) with LXCFS container-visible CPU counts; built with a no_std core, fail-open fallbacks, ABI-level tests, and fuzzing.

Keycloak — SPNEGO Mutual Authentication

Java · Merged upstream · 2026

Fixed final GSSAPI response-token propagation across Kerberos, LDAP, and HTTP authentication flows, with integration tests covering mutual authentication and transient credential state.

openssh-sftp-protocol — Non-UTF-8 Unix Filenames

Rust · Merged upstream · 2026

Fixed Serde Path decoding and request serialisation rejecting valid non-UTF-8 Unix filenames per the SFTP specification, uncovered via fuzz testing.

Smallstep — Issuer Validity Enforcement

Go · Open contribution · 2026

Proposed issuing-CA chain expiry clamping and warning propagation across X.509 issuance and renewal, with regression coverage and a companion CLI PR.

Technical Skills

Systems & Compute:
Linux, kernel debugging (vmcore, cgroups, PSI), CPU scheduling/NUMA, bare-metal infrastructure, NVIDIA H100/A100/H200, VMware vSphere/vSAN, storage I/O
Platforms & Cloud:
Kubernetes, Cilium/eBPF, Argo CD, Helm, AWS, Azure, Terraform, Ansible, Packer, Docker, TeamCity, GitLab CI, GitHub Actions, Nix
Networking & Identity:
BGP/anycast, TCP/IP, hybrid networking (ExpressRoute, Megaport), DNS, PKI (step-ca, ACME), Keycloak, Kerberos/SPNEGO, OIDC, Envoy Gateway
Data & Observability:
Kafka/Strimzi, PostgreSQL, Redis, VAST/S3, Trino, Hive, Prometheus, Thanos, Grafana, OpenTelemetry
AI Infrastructure:
NVIDIA HGX H200/H100, GPU scheduling (Sun Grid Engine), vLLM, LiteLLM
Programming:
Rust, Go, Python, Kotlin, Shell
Previously certified:
Red Hat RHCSA, Microsoft Azure AZ-700 (2024), AWS Certified Cloud Practitioner