BELLEVUE, WA — The ACM Seattle Chapter on Knowledge Discovery and Data Mining (KDD) hosted its latest monthly Summit on Friday, April 10, 2026, bringing together engineering leaders and experts for an evening of sharp technical discourse on the challenges of operationalizing artificial intelligence at enterprise scale. Held in Bellevue, Washington, the ACM Seattle KDD Summit featured a panel discussion with senior industry experts and two lightning talks showcasing novel approaches to DevOps automation and AI system optimization.
Attendees from across the Pacific Northwest’s thriving tech corridor gathered to engage with practitioners who are building, scaling, and rethinking the data systems that underpin modern tech advancements and moving the conversation beyond hype and into the architectural realities of production deployment.
A Panel That Cut to the Core of AI’s Infrastructure Gap
The evening’s centerpiece was a panel discussion led by Abhilash Kalwa, a Software Engineering expert, that brought together four distinguished voices in data science and engineering leadership, each offering a distinct lens on the obstacles standing between promising AI prototypes and reliable, scalable production systems.
Arun Vivek Supramaniana, Senior Data Engineer at Amazon, discussed some groundbreaking insights on the critical architectural components missing from the Modern Data Stack in the era of autonomous agents. His pioneering framework for managing the hidden costs of probabilistic workflows via cost-based semantic optimizers drew significant attention from the audience, demonstrating his extraordinary, industry-leading impact on AI-enabled data systems. As organizations rush to deploy agentic AI, Supramaniana argued, the underlying data infrastructure has not kept pace, and the economic consequences of that gap are only beginning to surface.
Shamnad Mohamed Shaffi, Senior Data Architect at Amazon, brought nearly two decades of industry experience spanning public safety, healthcare, and emergency response systems to the discussion. Shaffi spoke with authority about how data and AI are reshaping the way critical infrastructure operates today. “An emergency response system today isn’t just routing calls. It’s giving dispatchers context before they even ask, and the architecture underneath has to be fast, reliable, and compliant all at the same time,” Shaffi explained. He pointed to a persistent and troubling gap between AI usage and confidence, citing global research showing that while a majority of people use AI systems regularly, fewer than half actually trust them. Closing that gap, he argued, requires two things: transparency in how data is collected and used, and clear accountability across every layer of system ownership.
Manushi Sheth, Engineering Manager at Sonos, shifted the conversation toward the human side of the AI transformation. Sheth spoke candidly about where analysts should invest their energy as AI reshapes traditional data roles, and how to close the widening gap between what leadership envisions for AI and what teams can actually ship. Her remarks resonated with practitioners in the audience who are navigating the tension between executive ambition and engineering reality.
Rohit Nagpal, Senior Data Engineer at Amazon, rounded out the panel with a pragmatic message that cut through the noise: the journey from AI proof-of-concept to production isn’t about the model but it’s about the ecosystem around it. Infrastructure, governance, measurable ROI, and designing for scale from day one, Nagpal emphasized, are the factors that separate experiments from enterprise-grade AI systems.
Lightning Talks Showcase Next-Generation Engineering Concepts
Both lightning talks introduced novel technical concepts that underscored the Summit’s forward-looking ethos.
Alekhya Challa, Software Development Engineer at Expedia Group, presented a compelling vision for the future of DevOps through self-healing CI/CD pipelines. She highlighted how modern software delivery, while fast and iterative, is increasingly prone to deployment failures that lead to costly downtime and delayed recovery. Challa introduced self-healing pipelines as an evolution of traditional CI/CD, using closed-loop feedback systems to automatically detect, diagnose, and resolve issues without human intervention. By integrating monitoring, anomaly detection, and policy-driven automation, these pipelines significantly reduce Mean Time to Recovery (MTTR) and improve system reliability. A key innovation she discussed was temporal intelligence, the ability for systems to predict not just failures, but when they are likely to occur based on historical patterns. The talk emphasized the growing role of AI in enabling root cause analysis, automated patching, and adaptive system repair, allowing engineering teams to focus on innovation rather than firefighting operational issues.
Sravani Lingam, Data Science Manager at Infor, tackled a deceptively simple but costly problem: AI systems that repeat the same work when users ask similar questions in different ways. Lingam explained how this redundancy leads to higher costs and slower responses in real-world applications. She introduced the concept of semantic caching, a technique that helps systems recognize similar questions and reuse previous answers, reducing unnecessary computation and improving overall efficiency. Her talk demonstrated how semantic caching can significantly lower latency and cost while handling large-scale traffic, showing that sometimes the most impactful improvements to AI systems aren’t about building bigger models but about building smarter infrastructure around them.
A Growing Community Driving Knowledge Forward
The ACM Seattle KDD Summit is part of a monthly series of summits and tech events organized by the ACM Seattle Community on Knowledge Discovery and Data Mining. The chapter serves as a vital platform for tech professionals across the region to come together, share knowledge on recent developments in the data science and engineering world, and build the professional networks that fuel collaborative innovation.
As AI continues its rapid integration into enterprise systems, critical infrastructure, and consumer products, events like the KDD Summit play an increasingly important role, not just in disseminating technical knowledge, but in fostering the cross-disciplinary dialogue needed to build AI systems that are reliable, accountable, and ready for the real world.
The ACM Seattle KDD Chapter will continue hosting tech events, bringing together data and software engineers, data scientists, architects, and Engineering leaders who are shaping the next generation of intelligent systems in the Pacific Northwest and beyond.