Unlocking AI’s Future: QCon AI Boston 2026 Schedule Unveiled

The full schedule for QCon AI Boston 2026 is live! Scheduled for June 1-2 at Boston University, this two-day conference dives deep into the engineering challenges faced in deploying AI technologies in real-world applications. With a focus on transitioning from impressive demonstrations to production-ready systems, the program addresses critical topics like cost-effective inference, auditability in non-deterministic systems, and the evolving dynamics of software development when AI is part of the loop.

Bridging the Demo-To-Production Gap

Program chair Eder Ignatowicz, Senior Principal Software Engineer and Architect at Red Hat AI, highlights the essential dichotomy between a captivating AI demo and a system that can maintain stability and performance under real-world constraints. This focus sets the stage for sessions that explore the engineering hurdles that organizations must navigate to bring AI agents into production effectively.

Context Engineering for Agents

Agents typically excel during their testing phase, but their performance can falter when integrated into the intricacies of organizational services and data. This year’s conference will feature two key sessions focusing on context engineering to bridge that gap:

1. Context Engineering at LinkedIn
   Led by Ajay Prakash, Senior Staff Software Engineer at LinkedIn, this session explores the company’s implementation of the Model Context Protocol (MCP). The goal? Building an organizational context layer that tailors coding agents to work seamlessly with internal frameworks rather than applying a one-size-fits-all approach.

2. Beyond Prompting: Context Engineering for Production-Grade AI
   Ricardo Ferreira, Lead of Developer Relations at Redis, delves into the deeper aspects of building dependable AI applications. He discusses the necessity of gathering data and retrieving context to create outputs that maintain reliability in real-world applications, beyond simple prompt iteration.

Inference Economics and Infrastructure

For enterprises tackling AI on a grand scale, managing inference cost and latency is paramount. Three insightful sessions will investigate various facets of this critical issue:

1. Serving LLMs at Scale: The Hidden KV Cache Advantage
   Khawaja Shams, Co-Founder & CEO of Momento, presents research on how key-value caching can significantly enhance performance and reduce costs in serving large language models (LLMs). Discover the direct impacts on GPU utilization and throughput, as well as how to achieve lower “Time to First Token.”

2. Beyond the Prototype: Scaling Frame Agnostic AI Agent Infrastructure with Ray
   Apple’s own Deepak Chandramouli and Bhumik Thakkar discuss the evolution from prototype to a robust production-grade solution. Their session emphasizes the transition to an “Agent Engine” capable of handling large-scale web services effectively.

3. From Fab To Token: The State Of The Market
   Jordan Nanos, a Member of Technical Staff at SemiAnalysis, presents a detailed analysis of the physical and economic limitations currently constraining the AI infrastructure landscape, especially amid the contrasting strategies of traditional hyperscalers versus specialized “Neoclouds.”

Reliability, Evaluation, and Safety

The importance of safety, trust, and evaluation in AI systems cannot be overstated. Several sessions will tackle these topics head-on:

More Read

Optimizing Protein Engineering with Evolutionary Edit-Based Flow-Matching Techniques

Knowledge-Augmented Multimodal Clinical Rationale Generation for Disease Diagnosis Using Small Language Models: Insights from Paper 2411.07611

Scaling Discord’s ML Platform: From Single-GPU Workflows to a Shared Ray Cluster Setup

Claude Sonnet 4.5 Achieves SWE-Bench Verification and Expands Coding Focus to Over 30 Hours

Google Launches Project Suncatcher: Revolutionizing AI Models for Space Applications

• SafeChat: Building AI-Powered Safety Systems at Scale in a Real-Time Marketplace
  Bruna Pereira, Software Engineer at DoorDash, discusses how AI can ensure safety and trust in fast-paced marketplace interactions.

• Adaptive Recommenders in the Real World: Inference, Evals, and System Design
  Mallika Rao, Engineering Leader at Netflix, covers the continuous evolution of an adaptive recommendation engine, stressing the importance of real-time learning post-deployment.

• Building Reusable Evaluation Frameworks for Agentic AI Products
  Susan Chang, Principal Data Scientist at Elastic, shares methodologies for crafting evaluation frameworks tailored for agent systems that have been active in production for nearly two years.

• Zero Trust Agent Systems that Pass Audits and Still Ship
  Advait Patel, Senior Site Reliability Engineer at Broadcom, probes the challenges of deploying agent-based systems within stringent security regulations, emphasizing compliance without sacrificing performance.

AI Inside the Developer Workflow

As AI technologies continue to evolve, they’re reshaping the software development lifecycle (SDLC) and redefining engineering roles. The program will feature discussions highlighting these transformations:

• AI First, Quality Always: Agentic SDLC Adoption Case Study
  Catherine Weeks, Engineering Director at Red Hat, illustrates how to integrate AI-centric practices in the SDLC, maintaining a balance between productivity and reliability.

• Opening Keynote: The Five Stages of AI Maturity in Engineering Organizations
  Delivered by Lizzie Matusov, Co-Founder & CEO of Quotient, this keynote provides insight into common pitfalls organizations face on their journey toward AI maturity, as well as strategies to navigate these challenges.

Explore the complete schedule and secure early bird pricing or team discounts at boston.qcon.ai. Don’t miss this opportunity to gain invaluable insights into the forefront of AI engineering!

Inspired by: Source