Generative AI is rapidly evolving beyond prediction and prompts toward systems that can reason, act autonomously, and interact with the physical world. At the edge, this shift introduces new constraints and new opportunities, including low latency, energy efficiency, hardware diversity, privacy, and real-time decision-making.
The v5 Generative Edge AI Forum brings together industry and academic leaders to explore how reasoning, agentic, and physical AI can be designed, deployed, and scaled at the edge, and what it will take to move these systems from concept to production.
AGENDA DAY 1:
- Fireside chat; Bill Curtis – Moor Insights with Pete Bernard
Session 1: Industry Applications & Consumer Solutions
Focus: Real-world implementation, use cases, and consumer-facing edge AI.
- Juan Pisula (Roofline): Increasing innovation speed for agentic systems at the edge through heterogeneous AI compilation.
- Ashutosh Pandey (Infineon Technologies): Energy-Efficient, Multi-Modal AI through Edge Language Models (ELMs) for MCUs.
- Anirban Bhattacharjee (Wipro Limited): Audio Use Cases on edge devices (AI PC) – Real-Time and Offline Multilingual audio translation use cases.
- Sreenivasa Reddy Hulubeedu Reddy (was AT&T Services Inc): The 3.2% Hallucination Rate That Almost Killed Our AI — And How We Fixed It.
Session 2: Agentic AI & Distributed Systems
Focus: Agentic behaviors, Physical AI, and multi-agent distributed solutions.
- Danilo Pietro Pau (STMicroelectronics): Vision Language to Actions: A practical guide.
- Vladislav Sovrasov (Intel): Full-Stack Robotics VLA: A PyTorch-Native Path from Data Collection to Edge Deployment with ExecuTorch and OpenVINO.
- David Aronchick (Expanso): Zero-Trust Data Access for Agentic AI at the Edge.
- Gajendra Singh (Ambient Scientific): A Mixed-Signal Computing Paradigm for Physical & Agentic AI at the Edge.
