The 15 best podcasts for software engineers, ranked
The best podcasts for software engineers: system design, how real teams build and run software, engineering careers, and AI for software engineers, from AI coding tools to building with LLMs. Ranked for what a working engineer can use.
Gergely Orosz (writer of The Pragmatic Engineer newsletter; former engineering manager at Uber)
long-form interviews with experienced engineers · 75-120 min · weekly · since 2024
Why it is #1: The host asks for specifics about how named systems and teams actually work, so you hear real decisions and their costs. It covers classic engineering topics and AI coding tools with the same depth.
What it is like: Long conversations with engineers who built well-known systems, covering architecture, trade-offs and how their teams worked. Recent guests include Sam Newman, Peter Mattis, Casey Muratori and the Codex team at OpenAI.
SE Radio team (rotating volunteer hosts, including Robert Blumen, Gregory Kapfhammer and Priyanka Raghavan; run with IEEE Software, published by the IEEE Computer Society; founded by Markus Voelter)
one host interviews one expert on a single topic · 45-70 min · weekly · since 2006
Why it is #2: It aims to be a lasting educational resource rather than news, so most episodes still make sense years later. With over 740 episodes, there is likely one on the exact topic you are working on.
What it is like: Each episode takes one engineering topic, such as schema migrations, regression testing or agent state, and works through it with a specialist. It sounds like a structured technical interview, not a chat show.
swyx (Shawn Wang) (founder and CEO of Latent Space; co-hosts with Vibhu Sapra; Alessio Fanelli was the founding co-host)
interviews with AI engineers and founders, plus short Lightning episodes · 15-160 min · several episodes a week · since 2023
Why it is #3: It is the main show of the AI engineering field, so you hear early how builders use new models and tools. Guests are usually the engineers who built the product being discussed.
What it is like: Fast, technical interviews with people building AI models, agents, coding tools and infrastructure. Episodes vary a lot in length and topic, so you pick the ones that match your work.
Guy Podjarny (founder and CEO of Tessl, which runs the show; previously founded Snyk). Simon Maple (head of DevRel at Tessl)
interviews with engineers and engineering leaders about AI-assisted development · 30-65 min · weekly · since 2024
Why it is #4: It stays on the day-to-day practice of coding with agents: specs, review, testing and security. That makes it more directly useful for an engineer's own workflow than general AI news shows.
What it is like: Focused interviews about how teams write, review and ship code with AI agents. Guests are often practitioners such as Marc Brooker, Liz Fong-Jones and Dex Horthy rather than AI researchers.
Bring it questions like
Why do more AI-generated pull requests lead to more incidents?
Should I write specs before I let an agent write code?
How do I keep credentials away from my coding agent?
How much should I trust tests that an AI wrote?
Start with
Liz Fong-Jones: 2x the PRs, 1.5x the Incidents.What happens to reliability when AI doubles code output, and how to respond.
AWS's Marc Brooker: Specs, Not Code, Are the Hard Part.Why writing clear specifications matters more once agents write the code.
Best for: Engineers doing AI work: coding with agents day to day
Skip it if: You do not use AI coding tools and do not plan to.
Ron Minsky (engineer at Jane Street, the trading firm that produces the show)
long interviews with Jane Street engineers and outside researchers · 55-110 min · irregular, several episodes a year · since 2020
Why it is #5: The host and guests go further into the reasoning behind technical choices than almost any other show. Episodes are rare but each one is well prepared.
What it is like: Careful technical conversations about testing, build systems, networking, data systems and programming languages. Each episode has a full transcript on the show's website.
Kris Jenkins (independent developer and host; the show is supported by listeners on Patreon)
long interviews with developers who build databases, languages and tools · 60-140 min · about monthly, with some gaps · since 2023
Why it is #6: Guests are the people who wrote the tools, and the host lets them go into implementation detail. You come away understanding a tool well enough to judge whether to use it.
What it is like: Unhurried technical interviews where the guest explains how their tool works inside and why they built it that way. Topics include DuckDB, Turso, Jujutsu, eBPF and type-checking large codebases.
Adam Stacoviak (founder and editor-in-chief of Changelog Media; long-time co-host Jerod Santo retired from the show in March 2026)
long interviews with open source maintainers and developers · 95-165 min · every 2-6 weeks · since 2009
Why it is #7: It has followed open source for over 15 years and gets maintainers to explain the decisions behind their projects. Useful context for the tools your stack depends on.
What it is like: Long, relaxed interviews with the people behind open source projects and developer companies. Recent topics include Postgres, MCP, agents and supply-chain security.
Bryan Cantrill (co-founder of Oxide Computer Company, which runs the show). Adam Leventhal (engineer at Oxide Computer Company)
recorded live group discussion with Oxide engineers and guests · 60-90 min · most weeks · since 2021
Why it is #8: You hear experienced systems engineers explain how they found and fixed hard bugs in their own product. Few shows go this far into low-level debugging.
What it is like: A live Discord show where the hosts and Oxide engineers talk through debugging stories, Rust, hardware bring-up and computing history. It is loose and talkative, with real technical detail inside.
Hugo Bowne-Anderson (independent AI and data science consultant and educator; formerly at DataCamp, Outerbounds and Coiled)
long-form technical conversations with AI builders · 55-140 min · every 1-4 weeks · since 2022
Why it is #9: It goes into evals, verification and agent design in more depth than most AI shows. The host comes from data science, so methods are explained step by step.
What it is like: Long, careful conversations about agents, evals and how AI systems behave in production. Guests include Hamel Husain, Sebastian Raschka and the team behind the Amp coding agent.
Rotating hosts (including Sean Falconer (Head of AI at Confluent) and Gregor Vand (leads Ecosystem at Supabase); founded in 2015 by Jeff Meyerson)
technical interviews, plus a regular news episode · 45-65 min · twice a week · since 2015
Why it is #10: It covers a wide range of infrastructure and tooling topics twice a week, so it is a quick way to learn what a new category of tool does. Many guests come from the vendors that make those tools.
What it is like: One host interviews a guest, often a founder or engineer at an infrastructure or developer-tools company, about how their system works. Many episodes in 2026 are about AI coding and agents.
Jamison Dance (experienced software developer). Dave Smith (experienced software developer)
listener Q&A, two co-hosts · 30-35 min · weekly · since 2016
Why it is #11: It answers the non-technical problems engineers actually write in about, every week for ten years. Recent episodes deal directly with AI changing team habits.
What it is like: Two developers answer two listener questions per episode about work and careers, with a lot of jokes. The advice is practical despite the light tone.
Why can't I get promoted again so soon after my last promotion?
What do I do when coworkers flood code review with AI-generated code?
How do I handle inheriting a messy codebase when I already feel behind?
How do I mentor others without burning out?
Start with
Episode 534: I'm so bad at my job and I've inherited a huge slop pile.Handling self-doubt and a poor-quality codebase at the same time.
Episode 526: Why can't I get promoted so quickly after my last promotion and I got hired as a software engineer but ended up doing ops.How promotion timing works and what to do when your role drifts.
Best for: Individual contributors with career and workplace questions
Skip it if: You want technical content; this show is only about the people side of the job.
Wes Bos (web development educator who creates JavaScript, TypeScript, React and CSS courses). Scott Tolinski (creator of Level Up Tutorials; the show is produced by Sentry)
two co-hosts, plus occasional guests · 30-100 min · twice a week · since 2017
Why it is #12: It is the most consistent show for full-stack web developers, with over 1,000 episodes. You hear two working developers test tools and say what they think.
What it is like: Two web developers talk through frameworks, tooling, CSS and the AI tools they use, mixing tutorials with opinions. Friendly and fast-moving.
Daniel Whitenack (founder and CEO of Prediction Guard). Chris Benson (AI and autonomy research engineer at Lockheed Martin)
two co-hosts plus a guest · 45-55 min · weekly · since 2018
Why it is #13: It has run weekly since 2018 and explains AI deployment in plain terms. A good starting point for engineers moving into AI work.
What it is like: Two hosts and a guest talk through how AI is deployed in real organizations: agents, security, open models and enterprise rollouts. Explanations are plain and suit engineers new to AI work.
narrative episodes built around one story · 40-60 min · about monthly · since 2018
Why it is #14: Each episode is a carefully told story about real engineering decisions and their consequences. The lessons are indirect, but they stay with you.
What it is like: Produced story episodes about real software projects, bugs and careers, told like a radio documentary. Calm and easy to follow on a commute.
Ryan Donovan (host; the show is produced by Stack Overflow)
short interviews with engineers and tech company guests · 25-45 min · twice a week · since 2008
Why it is #15: Episodes are short and cover a new topic twice a week, so it is an easy way to keep up with how developers work. Some guests come from sponsoring companies.
What it is like: Short conversations about how software is built and run, with many recent episodes on agents, AI tooling and security. Easy to fit into a short commute.
Practical depth first: does an episode leave you with something you can use at work on Monday? Then the quality of the hosts and guests, whether the show is still publishing, and how consistent it is. Fame counts for little. Listener numbers are shown only when the show or a credible source has published them, with a link to where; we never estimate them. Facts were checked in November 2026.
Your question. Every show at once.
Have a burning business question you cannot find in one podcast episode? Describe it. TL;WL researches what founders and operators said about it across these shows and hundreds more, then narrates one private episode for you, with every quote linked to the minute it was said.
Free during the beta: 2 episodes a week of up to 15 minutes.
