Infrastructure and platform leadership.
What I was accountable for at EMBL-EBI, and what changed.
About 15PB to 150PB+ active
The active archive grew from about 15PB at ownership handover to more than 150PB, with about 400PB of raw storage underneath.
40 mount points to 3; 80 Globus installations to 3
Tickets fell from close to 100 a week to a handful a month. Effective throughput rose from 10 to 16Gbps to a sustained 60 to 75Gbps.
Tape DR tender, about £250K
The tender secured about £250K and cut DR total cost of ownership by 70%.
5,000+ federated users, six EMBL sites
Privileged administrators reduced from about 50 to 3.
8 direct reports
About 20 engineers influenced beyond the direct team.
About £1m a year
Annual infrastructure budget of about £1m, across an estate of 4,500+ Linux servers.
Infrastructure roles since 2007.
Deputy Head of Operations
Eight direct reports and about 20 engineers influenced; annual infrastructure budget of about £1m. Archive, transfer and identity services across an estate of 4,500+ Linux servers. Identity for 5,000+ federated users across six EMBL sites, with privileged administrators reduced from about 50 to 3.
Compute, then archive, transfer and identity services lead
Took on the active archive at about 15PB. Transfer services went from 40 mount points to 3 and from 80 virtualised Globus installations to 3. Wrote the tape DR tender that secured about £250K.
HPC systems engineer and trainer; technical lead, managed services; Technical Service Manager; Solutions Architect
Technical Service Manager from 2015; Solutions Architect in the final period, to September 2016. Bull was acquired by Atos in 2014, with the same colleagues and clients.
Research support coordinator and infrastructure lead
Ran the IT and research platform function as the organisation grew from about 50 to 250 people.
Full detail is in the CV: Download CV (PDF).
Chartered status, education and certification.
Chartered IT Professional, CITP MBCS
Awarded by BCS, the Chartered Institute for IT, in August 2026. BCS member since April 2026.
Google Professional Cloud Architect
Awarded September 2026.
Cambridge Judge Business School
CTO Programme, 2025.
Saïd Business School, Oxford
Executive Finance Programme, May 2026.
ITIL 4 Strategist
Direct, Plan and Improve, 2026.
PRINCE2 Agile
2020.
Red Hat, Oracle, NVIDIA
Red Hat RHCSA and RHCE, 2011, lapsed. Oracle Database 11g OCP, 2015. NVIDIA Deep Learning Institute, 2025: Fundamentals of Accelerated Data Science; Adding Knowledge to LLMs.
BSc Computer Science and Engineering
La Salle, Universitat Ramon Llull, Barcelona, 2005.
Foundations are permanent.
Architectures designed for a decade of operation, not for the next demo.
Side projects, built in my own time.
The same operating discipline at a smaller scale.
Everything I run at home is self-hosted, sits behind a single identity provider and a TLS-terminating edge, and is fully monitored. The discipline applied to a 150PB archive is applied here too: single sign-on, full observability, off-site replication, and documented runbooks.
- Public services behind one identity provider and a TLS edge
- Metrics and alerting across every host and service with Prometheus and Grafana
- Off-site ZFS replication with tested recovery runbooks
- A local-first AI inference stack: Ollama, LiteLLM, WhisperX, Qdrant
- Private git with CI on self-hosted runners
- Automation, documentation, and CRM on the same platform