I'm

I translate business problems into technical solutions - from requirements analysis and feasibility assessment through to proof-of-concept and delivery.

I've led technical teams as Team Leader across Solutions Delivery and Technical Services, balancing people leadership with hands-on delivery.

I've led PMO governance, change management, and vendor/contractor engagement across enterprise programmes - resourcing, financial forecasting, and risk/issue/dependency registers.

I build and manage relationships with stakeholders, clients and vendors, and gather business requirements through structured workshops and analysis.

I've coordinated proposal and RFP responses for enterprise clients, and trained stakeholders through workshops - from SAP finance training to onboarding support teams.

ITIL v4 Foundation certified - I apply structured service management practice to incident, change, and problem management.

I've developed and delivered SAP ECC finance and controlling training, building genuine working understanding of the platform for end users.

My core language for backend engineering - I reach for it when I need precise control over server-side logic, performance, and the Laravel ecosystem.

My framework of choice for building maintainable, well-tested backend systems quickly, without sacrificing structure or scalability.

I use Livewire and Blade to build reactive, server-driven UIs without needing a separate JS framework, keeping state and logic in one place.

The language of the browser - I use it for client-side interactivity alongside Alpine.js and Livewire.

The foundation of every interface I build - semantic markup and modern CSS, usually styled through Tailwind.

My utility-first CSS approach for building consistent, responsive interfaces quickly without fighting a separate stylesheet.

I've built interfaces with Bootstrap and Flowbite component libraries on projects outside my primary Tailwind/Laravel stack.

I've built with Flux UI and Flux UI Pro's component library for Livewire applications.

I've built and customised WordPress sites including WooCommerce stores and Elementor-built pages, alongside my primary Laravel work.

My dependency and build tooling for PHP and JS - installing, updating, and scripting the build pipeline.

I design and consume REST APIs, integrating third-party services and handling authentication, webhooks, and JSON data flows.

I build authentication and role/permission systems from scratch, including SaaS multi-user and subscription-based access models.

I design SaaS application architecture - multi-user, subscription/usage-based, and marketplace models - not just single-tenant apps.

I design and run headless CMS architecture, decoupling content management from presentation across multiple front-end surfaces.

I've integrated Stripe for payment and subscription billing within SaaS applications.

I build responsive, accessible interfaces as standard practice, not an afterthought.

My default relational database for structured application data - comfortable designing schemas, writing complex queries, and tuning performance.

I know and use both MariaDB and PostgreSQL in some of my Laravel applications, alongside MySQL.

I use Redis for queue and cache management in my applications.

I use S3-style object storage for images and larger files in my projects.

I design caching and queue-based background processing into application architecture rather than bolting it on later.

I integrate third-party systems via REST/JSON APIs and webhooks, and handle SSO/authentication integration between platforms.

I've been responsible for an enterprise web environment of around 26 public websites, covering governance, accessibility, and platform migrations.

Migrating and maintaining headless CMS platforms as they evolve.

Version control is second nature - branching strategies, rebasing, and history management for both solo projects and team delivery.

Managing dependencies and environment configuration - Composer, NPM, and .env files - across dev, staging, and production.

Provisioning and deploying servers and applications, from a fresh box through to a running production site.

I write feature and unit tests as part of the build, not an afterthought, using Pest for readable, expressive test suites.

I understand how large language models actually reason - token prediction, context windows, attention - not just how to prompt them. I design prompts and system instructions that constrain a model's behaviour deliberately, and I know when a task needs a bigger model, a smaller fine-tuned one, or no model at all.

I build and run my own AI infrastructure end to end, from a home lab GPU server through to cloud deployments, serving local LLMs with tools like llama.cpp and Ollama. I understand the trade-offs between self-hosted and hosted inference: cost, latency, privacy, and control.

I design multi-step agentic workflows that chain reasoning, tool calls, and external systems together, rather than treating AI as a single prompt-in, answer-out box. I scope an agent's permissions and failure modes so it fails safely rather than silently.

I build retrieval-augmented systems that ground model output in real data instead of relying on memorised training knowledge, and I know when that grounding actually matters versus when it's unnecessary overhead.

Beyond building with AI, I evaluate its output critically - spotting hallucinated facts, understanding a model's blind spots, and knowing when a deterministic script or a human decision is the better tool than an LLM.

Managing context windows and KV-cache state so inference stays efficient as conversations and sessions grow.

Serving models efficiently and benchmarking their real-world throughput and latency, not just their advertised specs.

I've built with hosted model APIs including Claude and Fireworks.ai alongside self-hosted Qwen-family models, routing between local and cloud inference depending on the task.

I apply OWASP principles and use OWASP ZAP for automated vulnerability scanning as part of my own development workflow. I'm currently working toward Security+ to formalise this further.

Implementing MFA, SSO, and identity and access management controls across the systems I manage.

Applying security policy and compliance requirements to the systems and processes I build and manage.

I've administered Microsoft 365 and SharePoint at scale - across an environment of 2,600+ sites - including governance, Teams, and Exchange mail flow.

OKTA, VMware, and MobileIron accredited - I've managed identity, access, and enterprise mobility platforms across a large device fleet.

Configuring Exchange mail flow and transport rules within a Microsoft 365 tenant.

I administer my own Ubuntu servers end to end - packages, services, permissions, and troubleshooting - for both production sites and my AI home lab.

Managing file and directory permissions, ownership, and ACLs on the servers I run.

I configure and run my own production web servers - nginx and php-fpm - rather than relying purely on managed platforms, so I understand the full request path.

I use CloudPanel to manage virtual hosts and sites on my own servers.

Issuing, installing, and renewing SSL/TLS certificates on the sites and servers I run.

Diagnosing server issues and managing disk/storage configuration on the infrastructure I run.

I build and administer my own AI servers, from disk partitioning and GPU passthrough through to running local model inference.

I manage DNS, domains, and Cloudflare configuration for production sites, including troubleshooting resolution and propagation issues.

The fundamentals underpinning every secure connection I deploy - certificates, ciphers, and the HTTP request/response cycle.

I configure and troubleshoot email authentication - SPF, DKIM, and DMARC - to fix mail-flow and deliverability issues.

Configuring LAN addressing and DHCP reservations across my home network and lab infrastructure.

Diagnosing connectivity and routing issues across LAN and server network configurations.

Supporting end users across Windows, macOS, Linux, and mobile devices.

Handling Tier 2/3 technical support and escalation management for enterprise users.

I've written CUDA C++ for GPU memory management - pinned and pageable host memory, asynchronous transfers, and streams/events for overlapping compute and transfer.

I've built model-memory streaming architecture for LLM inference, including bounded VRAM rings and host-memory-backed model weights that go beyond native VRAM capacity.

I've designed hot/warm/cold KV-cache tiering with eviction and scheduling to support concurrent inference sessions under memory pressure.

I've used profiling and byte-exact correctness testing to validate throughput improvements - including finding and eliminating an unnecessary weight copy-back that was silently capping performance.

Experimental research into FPGA/GPU workload partitioning for AI inference acceleration - PCIe-connected accelerators, DMA, and memory controllers - using a Xilinx Kintex-7 XC7K325T SmartNIC. This is research and prototyping, not production HDL development.

Understanding PCIe, DMA, and memory controller behaviour when connecting FPGA accelerators to a host system.

Exploring quantised inference formats and GEMV/GEMM operations for accelerated, lower-precision model execution.

I build and configure my own hardware - GPU selection, memory and PCIe topology, and BIOS/UEFI tuning - for both general use and AI workloads.

Configuring storage and memory - DDR, NVMe, and UEFI settings - when building and tuning my own hardware.

Diagnosing hardware faults, from component-level troubleshooting to full system builds.