AI Development Platform for the Defense Industry

Background: Software-Defined Defense

The defense industry is undergoing a major transformation from hardware-centric systems to software-defined defense. This shift toward advanced AI capabilities demands new tools and close collaboration between companies, research institutions, and government agencies. My client, Helsing, one of Europe’s most valuable startups, is at the forefront of this transformation.

Due to the confidential nature of the project, specific details cannot be shared.

Challenge: Secure and User-friendly Platform

Helsing and its partners set out to create a software platform for military AI development and evaluation—a space where multiple stakeholders across different organizations could collaborate effectively while maintaining the highest standards of data protection, security, and usability.

The key challenge was to design a unified user experience across a highly complex environment with diverse user needs, varying data classification levels, and multiple organizational and technical constraints (e.g. intellectual property concerns and strict network policy). The system needed to be intuitive for both technical and non-technical users, support complex AI development workflows, and enable fast iteration within a high-security environment.

My Contributions: From Strategy to Hands-On Execution

As part of the cross-functional delivery team, I contributed to all aspects of product design for the entire platform:

• Supported product management with backlog maintenance, feature prioritization, release planning, and coordination with partner teams.

• Led a small, multi-organization design team consisting of one freelancer and two agency designers.

• Conducted user research through surveys, in-person interviews, and usability testing, including System Usability Scale (SUS) benchmarking.

• Customized and maintained a comprehensive Figma design library based on MUI components to ensure UI consistency across the platform.

• Defined user flows, created wireframes, and crafted high-fidelity mockups to make complex workflows feel simple and intuitive.

• Collaborated closely with developers and meticulously tested the frontend to ensure high-quality implementation—fixing minor bugs and styling issues myself when needed.

• Authored and maintained user documentation to keep it clear, accurate, and user-friendly.

Outcome: Fast delivery with best-in-class UX

After just nine months, the first version of the platform was ready for users from ten different organizations. That’s remarkably fast for the defense industry where projects of this nature typically take years. The SaaS tool streamlined previously fragmented and manual MLOps workflows related to data sharing and transformation, AI model training, evaluation and auditing—all within a secure, collaborative, and user-friendly environment.

Following the initial release, the platform was iteratively improved based on user feedback, with new versions shipped every few weeks. The user experience was well received by stakeholders across organizations, and several users specifically praised the design and usability.

Results:

• 18% improvement in the SUS score after 4 month

• Strong positive feedback from users and stakeholders

Selected User Feedback: