Software Engineering Radio - the podcast for professional software developers
By SE Radio Team (team@se-radio.net)
Software Engineering Radio is a podcast targeted at the professional software developer. The goal is to be a lasting educational resource, not a newscast. SE Radio covers all topics software engineering. Episodes are either tutorials on a specific topic, or an interview with a well-known character from the software engineering world. All SE Radio episodes are original content — we do not record conferences or talks given in other venues. SE Radio is brought to you by the IEEE Computer Society and IEEE Software magazine.
45 episodes
All Episodes
Applying Systems Thinking to Diagnose and Mitigate Architectural Debt
Software systems fail when team structures conflict with architectural design. Master the principles of systems thinking to identify complexity debt, align your engineering team, and build a durable competitive advantage that AI cannot replicate.
View Episode Notes →
Integrating Accessibility as a Foundational Engineering Requirement
Accessibility is a core engineering requirement rather than an optional feature. When you treat it as a priority, you create clearer structures and better error handling. This shift turns fragile, exclusionary systems into resilient tools that work for everyone.
View Episode Notes →
Decoupling Storage and Compute via Apache Iceberg Architecture
Separate your storage from your compute to avoid vendor lock-in and stop expensive migrations. By using open table formats such as Apache Iceberg, you turn your data into a portable, version-controlled asset that makes your architecture more flexible.
View Episode Notes →
Building High-Fidelity Regression Harnesses Using Historical Data
Unit and integration tests often provide a false sense of security because they overlook the rare, high-impact production failures that occur in the wild. By replaying historical data at scale, you turn your testing process into a competitive advantage that builds genuine operational confidence.
View Episode Notes →
Securing Enterprise AI Through Modular Data Isolation and Guardrails
Centralizing data in massive vector stores creates a significant security risk for AI agents. You can protect your infrastructure by shifting to purpose-built, isolated data sources and implementing granular, identity-based access controls to prevent system-wide failures.
View Episode Notes →
Refactoring Engineering Workflows to Accommodate AI-Assisted Development
Deep technical expertise can become a liability when adopting AI because ingrained habits create friction. Leaders need to update their own mental models to avoid bottlenecks and prioritize operational results over how fast they can write code.
View Episode Notes →
Building Harness Engineering to Regulate AI Agent Behavior
Prompt engineering is a trap because it ignores how unstable AI agents actually are. Instead, focus on harness engineering. Build automated sensors and architectural guardrails to stop technical debt from piling up and to keep your systems stable over time.
View Episode Notes →
Why Probabilistic AI Models Require Continuous Behavioral Security Monitoring
Standard security measures do not protect AI systems because probabilistic models treat all input as instruction. You can protect your infrastructure by moving away from one-time build audits and toward continuous, systemic monitoring of model behavior and output.
View Episode Notes →
Prioritizing Deterministic Steering Over Fragile Agent Scaffolding
Complex AI agent scaffolding often creates fragility instead of reliability. By replacing rigid, over-engineered workflows with deterministic steering and just-in-time verification, developers build systems that are more robust and better able to adapt to rapid model innovation.
View Episode Notes →
Transitioning to Declarative Pipelines for Systemic Performance Gains
Library migrations fail when you treat them as simple syntax swaps. You can achieve massive performance gains by shifting from procedural, row-based code to declarative pipelines that use lazy evaluation and columnar architecture for long-term scalability.
View Episode Notes →
Why Design-First APIs Enable Parallel Development and Stability
Code-first development creates hidden bottlenecks that stall large projects. Adopting a design-first approach with upfront contracts decouples teams, enables parallel workflows, and builds the standardized, AI-ready infrastructure required for long-term system stability.
View Episode Notes →
Prioritizing Low-Latency Feedback for Real-Time Agentic Development
Modern Python development often suffers from slow feedback loops. By shifting from batch processing to real-time, incremental type checking, you can turn static analysis into an instantaneous tool for both human and AI-driven workflows.
View Episode Notes →
Relational Graphs End Information Loss from Manual Feature Engineering
The real bottleneck in predictive modeling isn't compute power, it's 30-year-old data preparation. Relational transformers treat databases as graphs, preserving structure for superhuman accuracy.
View Episode Notes →
How Trusted Intermediaries Enable Scalable Open Source
The Linux kernel scales through trusted intermediaries, not rules--a social architecture where delegation, not control, enables resilience. The real innovation isn't in code, but in how humans manage conflict and accountability at scale.
View Episode Notes →
Secrets Sprawl: Cloud, DevOps, and AI Amplify Systemic Vulnerabilities
Modern development tools amplify secret sprawl, creating systemic vulnerabilities and unprecedented attack surfaces. Discover how to manage these hidden costs before they become significant liabilities.
View Episode Notes →
Software Development Reimagined as Risk Management
Software development is risk management, not feature delivery. Understand the hidden risks behind methodologies like Scrum and Kanban to build more resilient software.
View Episode Notes →
Event Sourcing Reframes Software Development Beyond CRUD Illusions
Event sourcing reframes software development, revealing that upfront complexity leads to dramatically reduced costs and increased agility by treating every change as an immutable event.
View Episode Notes →
AI Agent Effectiveness Hinges on Context Control and Non-Interference
AI agents promise incident response autonomy, but true effectiveness demands meticulous context control and letting the AI drive the "how." This shifts focus from traditional frameworks to new validation needs.
View Episode Notes →
JavaScript Evolution: Deep Systems Understanding Over Surface Syntax
Unlock competitive advantage by mastering JavaScript's core systems. Go beyond syntax to understand the event loop and prototype chains, essential for building future-proof applications.
View Episode Notes →
Decoupling Observability: A BI Blueprint for Agility
Decouple your observability stack to escape vendor lock-in, slash costs, and gain agility. Embrace a layered architecture that mirrors BI evolution for ultimate flexibility and data access.
View Episode Notes →
Local-First Software: Control, Resilience, and Strategic Independence
Local-first software offers superior user experience and resilience by prioritizing local data control, moving beyond cloud-first dependencies and vendor lock-in.
View Episode Notes →
Contextual Awareness Transforms AI Ambiguity Into User Understanding
AI transforms frustrating interactions into seamless experiences by understanding context and user behavior, moving beyond error correction to embrace human communication's inherent ambiguity.
View Episode Notes →
Remote Pair Programming Requires Purpose-Built Tools for Low Latency
Remote pair programming thrives on sub-100ms latency. Discover how purpose-built tools unlock seamless collaboration and a competitive edge, far beyond generic conferencing solutions.
View Episode Notes →
Elm Architecture's Clarity Drives Robust Rust GUI Development
Embrace the Elm Architecture's separation of concerns in Rust's Iced GUI toolkit to drastically reduce debugging time and build more resilient applications.
View Episode Notes →
Sigstore Secures Software Supply Chain Through Cryptographic Provenance
Attackers exploit the software supply chain's hidden vulnerabilities. Discover how cryptographic links and transparency logs create trust, securing your development pipelines against evolving threats.
View Episode Notes →
Programmer's Judgment Drives AI Development Beyond Code Writing
AI transforms programming from writing code to guiding AI with judgment and taste, mastering an "ambiguity loop" for true control.
View Episode Notes →
Specification-Driven Development Unlocks AI-Powered Software Velocity
AI accelerates coding, but the real bottleneck shifts upstream. Invest in clear specifications to unlock sustainable velocity and high-quality software delivery.
View Episode Notes →
Vertical Integration Solves Data Center's Hidden Hardware and Software Costs
Vendor compromises create hidden data center costs. Understand why hyperscalers build their own hardware and gain strategic advantage by owning your entire tech stack.
View Episode Notes →
C's Deliberate Evolution and Nuanced Framework for Program Failure
C's slow evolution fosters stability, while a new framework categorizes program failures into four types, enabling more resilient system design and long-term maintainability.
View Episode Notes →
AI Transforms ERP: From Reactive Fixes to Proactive, Agentic Orchestration
AI transforms ERP from reactive fixes to proactive, data-driven foresight, demanding a strategic shift beyond automation to fundamentally reshape how businesses anticipate and manage operations.
View Episode Notes →
Hidden Costs of Observability Tool Migration Outweigh Perceived Benefits
Migrating observability tools incurs hidden costs, often exceeding benefits. Prioritize critical assets and validate every component for a resilient, cost-effective strategy.
View Episode Notes →
Continuous Architecture: Delay Decisions, Integrate Lifecycle, Align Teams
Delay decisions to build resilient, adaptable systems. Architecture must integrate build, test, and deploy, aligning teams with system design for long-term value.
View Episode Notes →
System Design Interviews Assess Maturity, Tradeoffs, and Communication
System design interviews assess your ability to navigate ambiguity, manage tradeoffs, and communicate complex solutions, revealing technical maturity beyond coding skills.
View Episode Notes →
Engineering Low-Latency AI for User Engagement and Retention
Achieve lightning-fast AI experiences by engineering latency out from the start, balancing accuracy, cost, and user engagement with proactive, critical-path optimization.
View Episode Notes →
Modern CLIs Evolve with Object Command Model and API Integration
Modern CLIs organize commands around objects and actions, integrating with APIs for distributed tasks and offering JSON output for interoperability, setting new standards for developer productivity.
View Episode Notes →
Requirements Maturation Flow: Shared Understanding, Bespoke Done, Ready
Stop project failure. Implement Requirements Maturation Flow to ensure shared understanding and clear readiness criteria *before* coding, reducing rework and stress.
View Episode Notes →
Layered Traffic Management and Bot Mitigation for Peak Demand
Manage extreme traffic spikes with virtual waiting rooms that ensure fairness, detect bots, and maintain uptime, preventing system collapse during high-demand events.
View Episode Notes →
Internal Developer Platforms Simplify Complexity and Reduce Cloud Waste
Internal Developer Platforms cut cloud waste by up to 45% and simplify complex multi-cloud environments, empowering developers with automation and controlled access to resources.
View Episode Notes →
Modular Agentic Architectures for Effective LLM Application Development
Build LLM applications with modular, agentic architectures, not monolithic models, for planning, orchestration, and specialized tools. Validate rigorously with multi-layered defenses.
View Episode Notes →
Building AI Native Software by Composing Multiple Models
Build AI-native software by composing multiple models into unique systems, moving beyond simple wrappers to achieve complex user goals and create differentiated products.
View Episode Notes →
Data Engineering Evolves: From Movement to AI-Ready Products
Data engineering transforms into a product-centric discipline, leveraging lakehouses and vector databases to power AI, embed governance, and create trusted, discoverable data products.
View Episode Notes →
Pragmatic Bookshelf: From LaTeX to Markdown for eBooks
Explore the intricate eBook infrastructure, from EPUB's dominance and Markdown's rise to the XSLT pipeline and the crucial role of human editors, all while preserving authorial integrity.
View Episode Notes →
Overture Maps: Unifying the World's Geospatial Data
Overture Maps Foundation liberates global map data into a unified, interoperable standard, empowering developers with cloud-native tools like GeoParquet and DuckDB for seamless integration.
View Episode Notes →
AI Debugging: Beyond Bug Fixing to Causal Understanding
AI dramatically accelerates debugging by sifting data and automating tasks, while time-travel debugging reveals causality, transforming "what happened" to "why," with human oversight remaining crucial.
View Episode Notes →
Prompt Injection: The Top LLM Security Threat for Enterprises
Prompt injection risks loom large for enterprise LLMs, allowing attackers to manipulate AI into leaking sensitive data or executing unintended actions, demanding layered defenses.
View Episode Notes →