STRADVISION, a vision recognition technology corporation developing AI for autonomous driving vehicles, presented its next-generation technology roadmap at AWS re:Invent 2025, held in Las Vegas, Nevada, U.S., from Dec. 1 to 5.
Kim In-su, head of STRADVISION's Data Innovation Center, gave the presentation, and STRADVISION introduced a hybrid on-premises–cloud integrated data architecture strategy to address growing complexity across data processing, curation and AI-based perception development by expanding collaboration with AWS (Amazon Web Services).
In his presentation, Kim In-su pointed to major challenges facing the global autonomous driving industry, including STRADVISION: the explosive increase in multimodal sensor data, model scale expansion and operational difficulties arising from expanded collaboration among global R&D organizations. He said that on-premises environments alone have begun to limit iteration speed, resource elasticity and the ability to run large-scale experiments, and that this has made it difficult to secure the core competencies needed to develop next-generation perception models.
Therefore, STRADVISION said it is modernizing its end-to-end data ecosystem by introducing a hybrid cloud pipeline that organically connects on-premises infrastructure and AWS.
The strategy allows the company to retain core internal systems while leveraging AWS for scalable data collection, processing, curation and orchestration capabilities. It also redesigned data workflows comprehensively but operates them without moving everything to the cloud, resolving bottlenecks and achieving more efficient, stable and resilient large-scale data operations.
Kim In-su said, "AWS provides the technical foundation to modernize the data ecosystem that is the core basis for autonomous driving development," and "by combining AWS in a hybrid structure, we have been able to operate large-scale data collection, curation, preprocessing and delivery processes flexibly, which was becoming increasingly difficult with an on-premises environment alone." He added, "This hybrid approach strengthens STRADVISION's platform, infrastructure and processes and is an important steppingstone that propels AI development to the next level."
Through the cloud-based innovation roadmap revealed this time, STRADVISION plans to continuously advance its next-generation data platform on a hybrid basis, including integrated global data accessibility, end-to-end automation from data ingestion to validation, and large-scale synthetic data generation. Through this, it plans to reliably strengthen efficient collaboration among global engineering organizations, rapid iterative development and consistent data quality.
[OSEN]