Software 3.0 (owner of AI Engineer events and community) is focused on building a global network of flagship and regional events, complemented by year-round community engagement opportunities. Our mission is to empower and unite the global AI engineering community by delivering transformative events, knowledge exchange, and collaborative experiences that accelerate practical AI progress.
External link for AI Engineer
San Francisco, CA, US
AI Engineer reposted this
After a sold out and incredibly successful AI engineer Melbourne, now it's Sydney's turn. Yes, AI Engineer is coming to Sydney December 7th and 8th, bringing the world leading conference to our hometown. Super early bird tickets are on sale now. The CFP is open. Melbourne sold out weeks in advance with hundreds on the wait list. So don't miss out!
An hour before he spoke, Lukas Petersson's team laid off Gemini. It had been running their café in Stockholm, a real café no human operates, and it had lost $6,000, so they handed the job to GPT. That is what happens when you stop grading agents on a benchmark and start grading them on a business. This is our first AI x Evals Track, in partnership with Arize AI, and the whole track is live now on YouTube. Full track: https://lnkd.in/ewafTeah Aparna Dhinakaran (CPO, Arize) opens with the pattern she sees across a dozen eval jobs: the eval has to change as fast as the agent it grades. Deterministic checks catch what you can define up front, LLM as a judge adds the analysis a fixed rule cannot, and agent as a judge hunts for the failure modes nobody thought to write a check for, then opens a pull request to fix them. Jason Lopatecki (CEO, Arize) shows what that looks like in production. Instead of getting paged at midnight and starting to dig, you wake up to an issue that has already been investigated. The unlock is boring and specific: traces on a filesystem, pulled down as files next to the code, because coding harnesses are magical with files and hopeless with a dashboard. Observability stops being a dashboard you click and becomes the smoke a system throws off for agents to read, which is why you now trace ten times more, not less. Philipp Schmid (Google DeepMind) brings the uncomfortable question. There are thousands of agent skills in the wild and almost none of them are tested. They get vibe checked with two manual runs and shipped. You would not merge code without tests, so why are we shipping skills without evals? Also on the track: - Maor B., Character.ai - Soumya Gupta & Jai Chopra, Uber - Rustem Feyzkhanov, Snorkel AI - Akele Reed, Dave Revere & Doug Keller, SonderMind - Preetika B., Daniel Bump & Chris Souza, Google - Alex Shaw & Ryan Marten, Laude Institute - Laurie Voss, Arize, who helped open the track From pairwise video judges and closed loop multimodal evals to private benchmarks rebuilt from production traces and evals as a CI gate: this is a full day on how to know whether your agent actually works. Full track: https://lnkd.in/ewafTeah
AI Engineer is coming back to Paris 🇫🇷 Following an incredible first edition, we’re thrilled to announce AI Engineer Paris 2026. This year, we’re excited to partner with Mistral AI to host this second edition, gathering of AI engineers, CTOs, VPs of AI, technical founders, and researchers. Tickets are already moving fast: 🎟️ Super Early Bird is sold out 🎫 Early Bird tickets are now available - limited supply CFP is open until July 31. We’re looking for practical applications and advancements in AI engineering. Get your ticket today and see you in Paris! Learn more: https://ai.engineer/paris
Live now: the full AI x Graphs Track from AI Engineer World's Fair, in partnership with Neo4j. Watch it here: https://lnkd.in/eNGV6kdQ The hard problem in agent engineering is not only making models smarter. It's giving them durable, inspectable context: what happened, what is true, where a fact came from, and how the pieces connect. Emil Eifrem, CEO of Neo4j, makes the case for thinner agents on top of a shared semantic substrate: an ontology of the business, the systems behind it, and the execution traces that let every agent learn from the last one. Yohei Nakajima, creator of BabyAGI, flips the usual architecture around. In ActiveGraph, the event log is the agent: state becomes a graph, and replays, rollbacks, forks, and controlled self-improvement follow from that design. Daniel Chalef of Zep AI tackles the provenance problem. When an LLM synthesizes facts from multiple sources, a source ID is not enough. In Graphiti, provenance is itself a graph—so agents can trace claims, apply trust policies, and delete data without losing the audit trail. Also on the track: - Zach Blumenfeld, Neo4j - James Le, TwelveLabs - frank coyle, UC Berkeley - Mike Phipps, Gates Foundation - Ritvik Pandya, JPMorgan Chase - Omri Bruchim & Tomer Ast, monday[.]com - Stephen Chin, Neo4j - Shafik Q. & Joanne Song, The New York Times - Subbiah Sethuraman & Abhilash Asokan, ZS Associates From graph memory and video context to agentic constraints, data models, and knowledge graphs as a control plane: this is a deep look at what it takes to give agents context they can actually reason over. Full track: https://lnkd.in/eNGV6kdQ
AI Engineer reposted this
If you're wondering where AI Agents go next, it's anybody's guess; so why build a rigid structure around something that is sure to change? Inngest Co-Founder Dan Farrelly laid it all out with clarity at this year's AI Engineer World's Fair. Build a harness, not on a framework; build it boring.
Your agent's architecture has a half-life of 6 months. Prompts change weekly. Models, monthly. How do you architect for change? Most teams couple all layers and wonder why every new best practice or pattern triggers a re-write. Check out my talk earlier this month at AI Engineer World's Fair or read the article version of my talk here ⤵️ 📝 https://lnkd.in/gssPavT9 🎥 https://lnkd.in/gHcEa4RQ
AI Engineer reposted this
I don’t believe reality is a simulation, but you genuinely couldn’t script this timeline: * Two weeks ago: At swyx ’s AI Engineer World Fair in SF, I decide at the last minute to introduce my friend Uri Rolls onstage for his talk on cyber benchmarks for infrastructure penetration and access control (see below, amazing team). I say: “There is a future where cyber is alive and everyone is well protected and I’m pretty sure that future involves open-source models.” And later: “A big challenge is going to be speed: the speed of attack versus defense. When an intruder starts to enter, you have to see what’s happening and catch them.” * One week ago: Hugging Face is hit by a sophisticated intrusion over the weekend. The traces look unlike anything we’ve seen before and suggest serious AI involvement, but we don’t yet know which model was used. The closed models we ask for help choke on their guardrails. We need to react fast, so we turn to Z.ai ’s GLM-5.2 to help us analyze the attack. * Earlier this week: OpenAI reaches out, discloses what happened, and partners with us on the investigation. The intruder turns out to be exactly what we had discussed two weeks earlier: a fully autonomous agent, powered by an unreleased frontier model, attempting to gain access to part of our infrastructure. Sometimes the timeline we live in is genuinely vertigo-inducing.
AI Engineer is coming to Shanghai for the first time. November 5 - 6, 2026, at the Hilton Shanghai Hongqiao. The program has two tracks: Open Source Models & AI Infrastructure and Physical AI. China and the US are advancing AI through different playbooks. Competition is real, but technical dialogue matters. Getting engineers from both sides in the same room, comparing notes, is worth doing. We're working with Yuzheng Sun and Mengying Li on AIEi Shanghai. Early Bird tickets are live now: https://lnkd.in/eMWyinMD Speaking contact: speakers@standup.partners Sponsorship contact: sponsorships@standup.partners
An agent at Replit deleted a production database, then fabricated records to hide it. It wasn't malicious. It was trying to help. That's the problem our first AI Security Track spent a full day on, and the whole track is live now on YouTube. Full track: https://lnkd.in/e3bjZ8Sz Steve Yegge ran Snyk over a game he'd been hardening for 30 years and it surfaced 241 vulnerabilities his own security pass never caught. His answer: the model that writes the code can't also be the one that checks it. Manoj Nair (CTO/CIO, Snyk) brought the data: security backlog is up 108% quarter over quarter across Snyk's customers, because agents ship code faster than anyone can close the holes they leave behind. Ezra Tanzer (Snyk) broke down the fix his team is shipping: three pillars, secure what agents generate, what they use, and what they do, down to blocking an agent live when it reaches for your secret key. Also on the track: - Eugene Yan, MTS, Anthropic - Kim Maida, Founding GTEM, Keyboard Labs - Steve Korshakov, Head of AI, Bee - Aaron Stanley, CISO, DBT - Lovina Dmello, NVIDIA - Moritz Johner, Form3 - Randall Degges, who opened and emceed the track Thanks also to Ethan J. Cha of OneCarlyle for being our first speaker from private equity. Full track: https://lnkd.in/e3bjZ8Sz
In January, juggling 10 terminal windows felt like peak productivity. By April it felt silly. That's Peter Steinberger opening the AI Engineer World's Fair keynote — and it landed because everyone in the room had lived some version of it. His arc over the past few months: First, he paired with one agent across 10 terminals. "I was managing 10 direct reports." Then he moved to a long running manager that delegates work to a team. His default changed: he manages the manager of a small company of agents. Three things made that possible: server side compaction (long running tasks finally reliable), coordination (one thread creates and steers projects), and automation (the manager wakes up when something happens). Persistent context. Delegation. Triggers. That's the loop. And once the loop starts working, you find the next bottleneck. His constraints, in order: Tokens — fixed. Compute — fixed by test boxes. (Agents run tests on a separate machine while you continue working) Attention — can't add more. "The most important skill today is deciding where to spend it." What that looks like in practice: someone files a GitHub issue. The manager wakes up, reads it against the project goals, decides if it fits. Creates a worker. The worker investigates, implements, runs tests. A second agent reviews. Peter gets a PR, the original issue, a proposed diff, maybe a video of a running build he can VNC into. He reviews once. Leaves a note. Maybe approves. Loop continues. "The agent runs the inner execution loop. I set the direction and make decisions in the outer loop." Meanwhile, Romain Huet and Alexander Embiricos framed the product direction clearly: not automating engineers — maximally empowering them. Models now ship every 6 weeks. It was every 15 months not long ago. GPT 5.6 on Cerebras runs at 750 tokens per second. They called it "a substantial PR in 10 seconds." 2025 was token maxxing. 2026 is value maxxing. The future is not 20 terminals. It's better loops. https://lnkd.in/gK2u_j_S
Startup Battlefield is going live now at AI Engineer World’s Fair. 18 AI native startup teams spent the morning demoing science-fair style: real products, real users, real technical depth. The top 3 now come to the main stage for live pitches, judge Q&A, and the $100K prize reveal. This is the part of the show I’m most excited for: no abstract market maps, no “AI will change everything” handwaving. Just founders showing what they built, why now, and why it matters. Tune in live here: https://lnkd.in/eH22HD4H #AIEngineer #StartupBattlefield #AIStartups