Saltmarch
Architecting for the Unknown

Architecting for the Unknown

21 Apr 2026, 08:45 GMT+05:30

In an environment of constant change and ongoing disruption, building systems that can adapt and endure is essential. This keynote explores the principles of adaptive architecture and how they enable enterprises to design systems that remain resilient while evolving with new demands and technologies. Attendees will gain insights into practical strategies for creating architectures that thrive under uncertainty and support long-term agility. What You Will Learn Core principles of adaptive architecture and system resilience How to design architectures that evolve with changing business and technology needs Practical strategies for building systems that remain stable amid uncertainty Who Should Attend Software architects, technical leads, engineering managers, and developers interested in resilient and future-ready system design.

Computer Programming is Dead; Long Live AI-First Programming

Computer Programming is Dead; Long Live AI-First Programming

21 Apr 2026, 09:45 GMT+05:30

The job market for computer science graduates is shifting rapidly, revealing a growing gap between academic training and industry expectations. Traditional programming education—focused on syntax and theory before practical application—no longer meets the needs of employers or students. In this keynote, the case is made that programming as we once knew it is effectively over. The future lies in AI-First programming, a new learning model built on a continuous cycle of trying, learning, and growing. Learners begin by building code with AI assistance, deepen understanding by asking AI to explain and refine that code, and expand their skills by testing and extending real-world applications. This approach accelerates confidence, builds practical capability, and develops the kind of AI engineers that modern organizations urgently need. What You Will Learn Why traditional programming education is failing to prepare graduates for modern software development How AI-First programming creates a faster, more applied path to mastery A structured loop of try, learn, and grow that builds confidence and real-world capability Who Should Attend Software developers, engineering leaders, educators and anyone interested in the evolution of programming education and the rise of AI-assisted development.

Beyond the AI Models: How Lowe’s is Building the Store That Knows

Beyond the AI Models: How Lowe’s is Building the Store That Knows

21 Apr 2026, 10:15 GMT+05:30

AI in retail is often discussed in terms of models and breakthroughs, but the real challenge lies in making it work on the store floor, in real time, for real customers and associates. In this keynote, Swaroop Shivaram shares how Lowe’s is using AI to transform how we shop, sell, and work, moving from experimentation to scaled impact. The session highlights two production solutions: Mylow Companion – a generative AI assistant that helps associates answer customer questions instantly. A computer vision system that detects where customers need help and enables proactive engagement. Beyond these use cases, the talk explores what it takes to operationalize AI at scale, engineering systems around models, ensuring accuracy and trust, managing hallucinations, and deploying computer vision systems at the edge. The session concludes with a perspective on how AI will redefine retail, turning stores into intelligent, assistive environments. What You Will Learn How Lowe’s has deployed generative AI and computer vision systems in production retail environments What it takes to operationalize AI at scale, including trust, accuracy, and edge deployment considerations How AI is transforming physical retail into responsive, assistive environments Who Should Attend Software developers and engineers Software and enterprise architects AI and machine learning engineers Platform and infrastructure engineers Technology leaders in retail and customer experience systems

Enterprise Architecture 2026–2028: AI-Native, Agentic, and Governed

Enterprise Architecture 2026–2028: AI-Native, Agentic, and Governed

21 Apr 2026, 11:00 GMT+05:30

Enterprise architecture is entering a new era defined by agentic AI, AI governance, confidential computing, and post-quantum cryptography (PQC), while sustainability and cost optimization are becoming architectural mandates. This session presents a practical operating model for architects to transform emerging technologies into trusted, scalable, and compliant platforms that meet CIO and CISO standards. Attendees will learn how to design an AI-native enterprise architecture: agentic workflows orchestrated with MCP and LangGraph, retrieval grounded in GraphRAG, governed under ISO/IEC 42001 and the NIST AI RMF, secured with OWASP LLM guardrails and confidential compute, and optimized through FinOps and GreenOps. The session also covers Software Carbon Intensity (SCI) metrics to measure cost and carbon per request, and strategies to prepare for PQC readiness using FIPS 203/204/205. It concludes with a 90-day activation plan and a three-year roadmap to modernize EA practices for the intelligent enterprise era. What You Will Learn Blueprint for designing AI-native, agentic enterprise architecture Governance alignment with ISO/IEC 42001 and NIST AI RMF GraphRAG and AgentOps patterns for explainability and resilience Security controls for LLMs, confidential compute, and PQC preparedness FinOps and GreenOps strategies with measurable ROI and SCI metrics Who Should Attend Enterprise and software architects, platform leads, AI program directors, and security or compliance leaders shaping the next generation of governed, scalable, and sustainable enterprise systems.

Breaking the Monolith: Tesco’s Journey to Federated GraphQL with xAPI

Breaking the Monolith: Tesco’s Journey to Federated GraphQL with xAPI

21 Apr 2026, 11:00 GMT+05:30

Tesco’s xAPI serves as the single entry point for all client interactions with the Retail Platform, powering web, mobile, in-store, and third-party experiences. Over time, this monolithic GraphQL API became a bottleneck, limiting scalability, capacity, and team autonomy. To address these constraints, Tesco evolved xAPI into a Federated GraphQL architecture, enabling independent subgraphs, dynamic schema composition, and domain-driven ownership. This session shares the practical journey from monolith to federation, including how the Strangler Pattern was applied for incremental migration, and how schema governance, observability, CI/CD pipelines, and multi-layer caching were implemented. The talk concludes with the measurable business and technical impact of federation at Tesco, including improved resilience and the convergence of store and online experiences under a unified API. What You Will Learn Why monolithic GraphQL APIs become bottlenecks at scale How to apply the Strangler and Modular Monolith patterns to migrate safely to a federated architecture The business and technical impact of GraphQL federation within a large retail platform Who Should Attend Backend developers API engineers Software architects Platform and infrastructure engineers Engineering leads responsible for API scalability and modernization

A Practical Introduction to LangChain4j

A Practical Introduction to LangChain4j

21 Apr 2026, 11:00 GMT+05:30

Interested in adding AI capabilities to your Java applications? LangChain4j makes it simple to integrate large language models (LLMs) directly into your existing codebase without leaving the Java ecosystem. In this session, we will go beyond “Hello World” examples and explore real-world use cases where AI can enhance everyday applications, from intelligent assistants and document summarizers to data enrichment and natural language interfaces. Through live coding and practical demos, you will learn how to connect to models, chain prompts, manage context, and integrate AI workflows into Spring or Micronaut applications. By the end, you will have a clear understanding of how to design, implement, and extend AI-powered features in Java using LangChain4j. What You Will Learn How LangChain4j bridges Java and large language models Practical examples of integrating AI features into real applications How to build, chain, and orchestrate AI prompts and tools Best practices for managing context, inputs, and outputs How to extend LangChain4j with custom tools and connectors Who Should Attend Java developers and architects looking to apply AI practically in production applications using familiar frameworks and tools.

The AI-Native Codebase

The AI-Native Codebase

21 Apr 2026, 11:00 GMT+05:30

AI coding agents are now a default part of everyday software development, yet many teams struggle to use them reliably at scale. While AI can generate code quickly and in large volumes, that output often becomes difficult to review, understand, and maintain over time. As a result, adoption is frequently driven by trial and error rather than by deliberate design. This session presents a five-level codebase maturity framework for creating and evolving codebases that support sustainable, production-quality development with AI coding agents. Each level defines clear goals, checklists, assessments, and success criteria, all grounded in real-world case studies. The talk explores how this framework leverages AI strengths such as speed and pattern recognition, while addressing weaknesses related to correctness, context loss, and long-term maintainability. The focus is on enabling effective human and AI collaboration so teams can ship reliable software at scale. What You Wwill Learn A five-level maturity framework for assessing and evolving AI-ready codebases Practical criteria, checklists, and success measures for each maturity level How to balance AI-generated code with human oversight to maintain production quality Who Should Attend Software Developers Software Architects Technical Leads and Engineering Managers Teams adopting or scaling AI-assisted development

AI-Powered MongoDB ETL Without the Pain

AI-Powered MongoDB ETL Without the Pain

21 Apr 2026, 11:00 GMT+05:30

Building reliable ETL pipelines for MongoDB requires balancing speed, complexity, and governance. As data volumes grow and use cases expand across analytics and compliance, traditional approaches can become brittle and time-consuming. This session explores how AI-assisted techniques are reshaping MongoDB ETL design, using real-world scenarios to demonstrate practical approaches. The talk covers how natural-language-driven pipeline creation, automated transformations, and unified workflows can simplify common challenges such as data masking, aggregation for analytics, and event streaming with Kafka. It focuses on modern ETL patterns that reduce operational friction, shorten development cycles, and make MongoDB data pipelines easier to build, evolve, and govern. What You Will Learn How to build MongoDB ETL pipelines using natural language with AI-generated transformations How to handle real-world use cases such as data masking, analytics aggregation, and Kafka-based event streaming How AI-assisted workflows can reduce pipeline development time and operational complexity Who Should Attend Data engineers Backend developers working with MongoDB Platform and infrastructure engineers Architects designing data pipelines Teams working on analytics, compliance, or event-driven systems

AI Is Not the Risk. Architectural Drift Is

AI Is Not the Risk. Architectural Drift Is

21 Apr 2026, 11:30 GMT+05:30

AI is accelerating how quickly software is written and delivered. As delivery speeds increase, systems begin to evolve faster than they are designed. The primary risk is not incorrect code, but the gradual erosion of system integrity through inconsistent decisions, hidden coupling, and the loss of clear architectural boundaries. This session examines how engineering teams can sustain delivery velocity while maintaining control over system design. It focuses on identifying early signs of architectural drift and adopting practices that preserve structure, clarity, and long-term system integrity as development accelerates. What You Will Learn Why architectural drift is a greater risk than incorrect code in AI-assisted development How system integrity degrades through inconsistent decisions and hidden coupling Approaches to maintain architectural clarity while sustaining high delivery velocity Who Should Attend Software developers Software architects Platform engineers Engineering managers and team leads Teams working with AI-assisted development workflows

AIOps: Leveraging AI for Incident Root Cause Analysis

AIOps: Leveraging AI for Incident Root Cause Analysis

21 Apr 2026, 11:30 GMT+05:30

This session focuses on how agentic AI is used within the Jira Service Management (JSM) ITOps product at Atlassian to assist engineers with root cause analysis during production incidents. It examines the role of context engineering in enabling effective RCA, and how service topology graphs and DevOps context graphs, powered by the Teamwork Graph, support this process. The session also shows how the RCA agent combines signals from sources such as Bitbucket or GitHub pull requests, feature flag rollouts, MELT and observability platforms, and knowledge bases to triangulate and identify the root cause. It will also discuss the trade-offs between a multi-agent approach and a skills-based approach for effective agent design. What You Will Learn How agentic AI supports root cause analysis in production incidents The role of context engineering and graph-based systems in RCA Trade-offs between multi-agent and skills-based approaches for agent design Who Should Attend Developers Site Reliability Engineers DevOps Engineers Platform Engineers Architects Engineering Managers