Multiverse Computing
Solutions Architect, Nov 2025 to July 2026, Madrid. I ran technical pre-sales for CompactifAI, Multiverse Computing's LLM compression product: discovery, demos, proofs of concept, and deployment plans.
The company
Multiverse started in San Sebastián with quantum-inspired algorithms and moved into AI model compression. CompactifAI uses tensor networks to shrink large language models by up to 95% while keeping their performance, so they run on less hardware at a lower cost.
- June 2025. $215M Series B led by Bullhound Capital, with HP Tech Ventures, Toshiba, Forgepoint, and Santander among the investors. I joined a few months later.
- July 2026. Series C targeting up to $570M at a $1.7B pre-money valuation, about five times the Series B. It was co-led by Forgepoint, BNPP Solar Impulse Venture Fund, and Bullhound, and announced the month my role ended.
Pre-sales
A discovery session usually started with three inputs: the model the customer wanted to run, the hardware they had, and how much quality they could give up. I compared memory use, latency, inference cost, and quality loss, then wrote a recommendation an engineer could test and an executive could price.
I built and presented the demos for customer meetings, webinars, and events. The same result had to make sense to an ML engineer asking about benchmarks and to a manager asking what it would cost per month.
Most deals involved engineering, infrastructure, procurement, and leadership, each with its own concern. A good part of the job was getting them to one decision and writing it down as a reference architecture.
Deployments
Deployment planning ran with AWS, NVIDIA, and Dell, for both cloud and on-premises setups. I contributed to an on-premises GPU deployment for a defense-sector customer.
Sales tooling
I built reusable AI workflows for sales preparation and technical proposals, helping the sales and presales team work from consistent materials.