Engineering the Future of AI by Endeavor скачать в хорошем качестве

Engineering the Future of AI by Endeavor

Speakers: • Panos Stravopodis, Co-founder & CTO, Elyos AI (Voice AI for field services) • Dimi Ilias, Co-founder & CTO, Intryc (Automating QA for customer support) • Petros Mandalis, Co-founder & CTO, Proxyfood (AI for food recipe formulation) • Moderator: Niko Efstathiou, Editor in Chief, WIRED #CTOs #EngineeringAI #VoiceAI #ProductionAI #AIEngineering #TechnicalLeadership A technical roundtable featuring three CTOs discussing the transition from prototype to production. They explore the specific engineering challenges of their respective fields—voice latency, food science data complexity, and QA at scale—and share insights on building reliable systems that move beyond weekend demos to enterprise-grade products. Key Highlights: ✅ Voice AI Engineering Challenges: Real-time latency is critical—human conversation happens in ~400ms turns but current AI stacks are often closer to 1 second. Emotional intelligence is essential for detecting if a caller is panicked and adjusting tone accordingly. Conversations are modeled as "state machines" rather than open-ended chats to ensure consistency. ✅ Food Science AI Complexity: The challenge is combining unstructured scientific data with predictive analytics to create recipes that are chemically viable and reproducible in a lab, not just AI hallucinations that look plausible but fail in practice. ✅ QA at Scale Architecture: When processing millions of support tickets, human review is impossible. The solution is an "AI watching AI" architecture, where specialized models evaluate the performance of customer-facing agents automatically. ✅ The "Junior Engineer" Crisis: Concern about the shrinking entry-level market—if AI agents do the junior work, how does the next generation of senior engineers learn the foundational skills and gain experience? ✅ Hybrid Evaluation Requirements: For high-stakes decisions (healthcare, banking), a "human-in-the-loop" is non-negotiable. Systems should be designed to escalate low-confidence predictions to humans automatically rather than forcing binary automation. 💬 Why Watch? Critical for CTOs, engineering leaders, and technical founders. This panel reveals the massive engineering gap between demos and production systems. Learn about the specific technical challenges of voice AI, domain-specific AI, and quality assurance at scale from teams actually building these systems. 📌 Notable Insights: "Vibe coding" skepticism from CTOs—while useful for rapid prototyping, robust, maintainable backend code still requires skilled engineers who understand architecture. The biggest barrier isn't technology—it's the "emotional rollercoaster of a founder" and maintaining resilience through the inevitable challenges of building. "If you don't have good data, you just have a faster way to make mistakes"—data quality is the foundation that determines whether AI adds value or amplifies errors. Production AI requires rigorous, metric-driven evaluation frameworks (like RAGAS) rather than "vibe checks" to ensure consistent performance at scale. About the GenAI Summit: Over 5,000 GenAI builders, technology executives, and political and business leaders convened on November 24 at the Stavros Niarchos Foundation Cultural Center for the 4th GenAI Summit, organized by 100mentors. Under the theme "From Chatbots to Agents," the Summit captured the evolution of GenAI: from simple Q&A systems to autonomous platforms that plan and execute complex tasks. European product "premieres" and participation from leading international companies cemented the GenAI Summit's position as Southeast Europe's most influential artificial intelligence conference. Founded in 2023, GenAI Summit has established itself as Southeast Europe's most influential AI event. It convenes researchers, entrepreneurs, and government leaders shaping GenAI's future, connecting scientific research with practical implementation across enterprises and organizations. For more information: genaisummitseeurope.com #CTOs #EngineeringAI #VoiceAI #ProductionAI #AIEngineering #TechnicalLeadership #GenAISummit #Innovation #Technology #ArtificialIntelligence