Services

How I can help

AI and data architecture

For organisations adopting AI that need an architecture which survives contact with production. I design data and AI systems around what the business actually needs to measure.

Typical deliverables. Architecture review and target-state design; platform and vendor evaluation against scored criteria; proof-of-concept design with success metrics agreed before the work starts; production-readiness and data-quality assessment; governance and data-trust design.

Relevant experience. Since 2019, enterprise AI and data platforms — MathWorks, DataRobot, Dataloop, Anomalo — leading multi-stakeholder technical evaluations for global enterprises across energy, automotive, defence, insurance, telecoms and retail.

HPC and simulation performance

For teams whose numerical codes are too slow, too expensive, or both. Profiling, parallelisation and acceleration.

Typical deliverables. Performance profile and bottleneck analysis; parallelisation and GPU roadmap; feasibility assessment for ML surrogates and learned preconditioners, qualified on kernel share rather than on enthusiasm; implementation and benchmarking against a defined baseline.

Relevant experience. PhD in computational physics; postdoctoral Monaco Fellow at the ITER Organization; author of PB3D, a parallel Fortran code for three-dimensional magnetohydrodynamic stability; lecturer in plasma physics and computational methods for HPC at Eindhoven University of Technology.

Technical go-to-market for deep tech

For research-heavy companies with excellent technology and a need for professional commercial motion. Who exactly to sell to, how to qualify them, and how to run pilots that convert rather than merely impress.

Typical deliverables. Ideal-customer profile and qualification criteria; positioning and technical narrative; pilot and proof-of-value playbook; pipeline review and forecast discipline.

Relevant experience. Led a global pre-sales solutions architecture team; designed the methodology for full-cycle proofs of concept and enterprise evaluations; built the technical narrative and enablement that shortened enterprise sales cycles. Enterprise qualification frameworks (MEDDICC) and consultative technical selling, applied through procurement, security and data-governance review at global enterprises.

Training and workshops

For technical teams, and for executives who have to make decisions about technology they do not build themselves.

Typical formats. Half-day and full-day workshops with hands-on labs; executive briefings; short lecture series. Topics include HPC acceleration, LLM fine-tuning and evaluation, agentic systems and their failure modes, and pragmatic AI adoption.

Relevant experience. University lecturer at TU/e in the BSc Applied Physics and MSc Nuclear Fusion programmes; public workshops and webinars across seven years in technical vendor roles.

How I work

Every engagement is scoped to named deliverables with a fixed end date. You always know what you are getting.

• Diagnostic: two to five days, ending in a written assessment you can act on whether or not we work together afterwards.
• Project: defined deliverables, agreed timeline, fixed end date.
• Advisory: typically one to two days per week, reviewed each quarter.

Remote or on-site. Working languages: English, Dutch, Spanish (selected topics).

Rates on request, and I will tell you early if I think the work does not need me.

Get in touch

Best reached by email at info@toonweyens.com.

Also on LinkedIn, GitHub for code, and Google Scholar for publications.

The full curriculum vitae has the detail.