Experience
I’ve worked across data science, applied AI, and the infrastructure behind it. Today, I build and run LLM inference platforms on Kubernetes.
- Python
- LLM inference
- vLLM / NVIDIA Dynamo
- Kubernetes
- Platform engineering
- AWS
- Nix
- Linux
Likezero
Sep 2024 — PresentSenior AI Engineer
- I built and own Likezero’s AWS Kubernetes LLM inference platform, using vLLM and NVIDIA Dynamo with disaggregated prefill/decode, KV-cache-aware routing, speculative decoding, and GPU autoscaling through Karpenter and KEDA.
- I optimise cold starts, KV cache reuse, and decode throughput for Llama, Qwen, DeepSeek, and embedding models on RTX 6000 Blackwell, H100, and H200 GPUs.
- I run the platform through declarative GitOps with Argo CD and Terraform, Cilium networking, cert-manager for TLS, and Prometheus, Loki, and Grafana for observability.
- I built a high-concurrency asynchronous Python server bridging RabbitMQ and HTTP, enabling decoupled, KEDA-driven autoscaling. It has handled hundreds of millions of LLM and internal service requests.
Applied Data Science Partners
Aug 2023 — Sep 2024Senior AI Engineer
- I delivered three LLM projects in defence and education, working with GPT, Gemma, and Llama, alongside RAG and retrieval-augmented reasoning.
- I led a generative AI and software engineering project applying decoder transformers to new use cases in cyber defence.
- I wrote an AI-centred newsletter that grew to more than 850 subscribers and supported new consulting opportunities.
Liveminds and Maeving
Jul 2023 — Sep 2024AI & Data Engineer Consultant
- At Liveminds, I productionised a text summarisation product using multi-step RAG with GPT and Claude, helping launch the company’s AI offering and attract investor interest.
- At Maeving, I built an ETL and analytics pipeline with Azure, Postgres, and Power BI, giving teams a trusted shared source for self-service analytics.
Laing O’Rourke
Jun 2021 — Jul 2023Data Scientist
- I delivered an end-to-end Azure ML data product for multi-billion-pound financial data, supporting economic strategy for national infrastructure projects.
- I fine-tuned Sentence BERT to understand construction-specific language, opening up new NLP use cases across the industry.
- I taught Python and data analytics skills and shared developments in data science across the company.
FACEIT
Sep 2020 — Jun 2021Junior Data Scientist
- I analysed matches from more than two million players to build skill prediction models and address smurfing.
- I improved toxicity detection across millions of monthly chat messages, improving the platform experience for millions of users.
- I built dashboards and data marts with Looker and BigQuery to support decisions across the business.
Various companies
Oct 2018 — Aug 2020Freelance Data Scientist
- I built and deployed a predictive model that reduced costs associated with advertising and employee time by 50%.
- I delivered business intelligence across several companies and productionised a system to scrape, clean, and store NHS jobs data.
Education
Keele University
Sep 2016 — Jul 2019BSc (Hons) Physics and Mathematics