Vijay Kakani
Vijay Kakani is a Computer Vision Engineering Manager at Rivian Volkswagen Group Technologies, where he leads a cross-functional team developing edge perception and multimodal AI systems for automotive-grade hardware. He holds a PhD in Information and Communication Engineering, specializing in Future Vehicle Engineering, serves as Adjunct Faculty at Inha University in South Korea, and was named to Stanford’s World’s Top 2% Scientists list in 2025, with more than 60 SCI-indexed publications. His work focuses on the hardware, software, and team architectures needed to make embedded AI repeatable, safe, and scalable across a global fleet.
Talks
Intelligence at the Edge: Deploying Multimodal AI Under Extreme Compute Constraints
The next generation of vehicles will not just perceive the road; they will rely on multimodal AI to understand context and support drivers in real time. Bringing these capabilities into a production vehicle is not simply a matter of selecting a larger model; it requires working within strict memory, latency, power, thermal, and reliability limits, including operation without connectivity. This talk explores the engineering realities of deploying computer vision and multimodal AI directly on embedded automotive hardware, covering model selection, quantization, optimization, and the validation discipline required to move an AI capability from research demo to a dependable, offline-capable vehicle platform.