🇫🇷 Watch along with us on AI Engineer Paris 2026: piped.video/watch?v=CGq9KRSb… Opening Keynotes with: - @MistralAI ! - @liamcbride! - @rawert! - @thekitze! The next level of AI engineering, direct from the beautiful @joinstationf in Paris!

really enjoyed the more intimate vibes of @aiDotEngineer Paris yesterday and catching with old friends, s/o the Mistral team and @AlisdairBroshar !!

AI Engineer Paris 2026 is over. Great talks at @joinstationf, and even better hallway conversations. So many sharp people from the AI world in one place. Thanks @MistralAI and @aiDotEngineer 🤝

Yesterday, I attended the @aiDotEngineer event at @joinstationf in Paris. The event brought together AI builders, founders and engineers from companies such as @MistralAI, @nvidia, @neo4j, @sentry, @liquidai, @stripe, @huggingface, @ElevenLabs, @bfl_ai, @modal, @Cloudflare, @CockroachDB, Daytona, @qdrant_engine, @langfuse, @FactoryAI and many others. The room was full of impressive demos and ambitious ideas. But one question kept coming back to me: What happens after the prototype works? That is where real AI engineering begins. The lesson I took away is simple: Building an AI demo is exciting. Building an AI system that is reliable, secure, observable, cost-effective and useful in production is the real challenge. For engineers building AI products, I believe the priorities should be: 1. Start with a real business problem, not just a model. 2. Define evaluation criteria before launching. 3. Build strong data and knowledge foundations. 4. Monitor quality, latency, cost and failure cases. 5. Keep humans in the loop where decisions matter. 6. Treat security, privacy and governance as engineering requirements. The future of AI will not be defined only by who builds the most powerful models. It will also be defined by who can turn those models into dependable systems that create measurable value. What is the biggest challenge you face when moving an AI project from prototype to production? A special thank you to @MistralAI for providing the tickets and making it possible for me to attend this inspiring event. Follow me @sugumaran___ for more insights on AI engineering, agentic systems, MLOps and building reliable AI solutions.

Looking forward to speaking at @aiDotEngineer New York, Oct 12–14. Come say hi — ai.engineer/nyc/2026

AGI will not be typed, it will be spoken: that's the closing line of "Speech-to-Speech Model Research at Google DeepMind," the talk @valeriawu_ and Tom Ouyang, both of Google DeepMind, gave, and @aiDotEngineer has posted it on YouTube. It traces how Gemini's natively multimodal pretraining replaced the old pipeline of separate acoustic, pronunciation, and language models, and shows what one speech-to-speech model can do once translation, tool use, and conversation all run through it. - Live translation. Gemini's live model does real-time speech translation across 70+ languages, preserving each speaker's voice, with streaming quality that holds up against offline systems that get to hear the whole utterance first. - Three things pulling against each other. The talk frames the model's goals as conversational quality, intelligence, and multimodality: push "thinking" higher and evals show more intelligence, but latency and naturalness take the hit. - Knowing when to stay quiet. A feature called proactive audio lets the model decide not to respond when it hears background noise or someone else talking, since most real conversations aren't happening in a quiet room. - Non-English first. Most Gemini users aren't English speakers, so translation and localization work covers all their languages, not just English. - Two very different demos, one model. A Search Live clip identifies a boucle sofa in Spanish, localized to Spain's Spanish, correctly leaving "mid-century" in English; a Live API demo has the same model handling a roadside assistance call, reading back a car's registration plate and postcode. - Faces, not just voices. A pilot with Citi shown at Cloud Next adds real-time, multilingual, lip-synced avatars on top of the same model. I'm working through the published talks from AI Engineer World's Fair sharing summaries and takeaways. Follow for more!

thanks to everyone who joined my session at @aiDotEngineer paris yesterday! for everyone who missed it, here's the thread form of my talk 🧵

Looking forward to speaking at @aiDotEngineer New York, Oct 12-14. Come say hi! ai.engineer/nyc/2026

Retour sur le dîner des speakers @aiDotEngineer mercredi soir. Ravis d'avoir partagé ce moment avec nos amis et partenaires 😊

AIE Code Summit CFP is now open! Talks, workshops, and keynotes: we want to hear from people building AI coding tools and actually shipping them. What worked, what broke, and what you'd do differently. Submit by October 11th ➝ ai.engineer/cfp/code

Announcing @tariqshaukat, CEO of @SonarSource, at AI Engineer New York. His session, “The Machine that Changed the World, vAI,” lays out a verification-first framework for agentic software factories that can ship quickly without sacrificing trust. October 12–14 → ai.engineer/nyc/2026

Announcing @MrMichaelADavis, Global Chief Security Architect at @jpmorgan, at AI Engineer New York. His session, “Intent Based Auth,” examines why governing an agent needs more than allow-or-deny tool permissions: its actions need to be evaluated against the mission it is trying to accomplish. October 12–14 → ai.engineer/nyc/2026

Announcing @ryanlstevens12, Director of Applied Science at @tryramp, at AI Engineer New York. His session, “We gave our finance agent more context. It got worse,” shares Ramp’s eval-driven approach to context, skills, memory, and agent-harness design—and what actually improves quality, reliability, latency, and cost. October 12–14 → ai.engineer/nyc/2026

Announcing Sam Green, Team Lead, Applied AI - AIA Labs at Bridgewater, at AI Engineer New York. His session, “Vault: When Fifty Years of Structured Data Isn't Enough,” explores how Bridgewater built a shared unstructured-data platform that powers investment research and production AI agents. October 12–14 → ai.engineer/nyc/2026

Announcing Aaron Linsky, Head of Product Engineering at Bridgewater Associates, at AI Engineer New York. His session, “Can Your AI Analyst Keep a Secret?,” goes inside the retrieval layer of a production AI analyst—showing how Bridgewater and turbopuffer enforce changing, row-level permissions over millions of confidential documents. October 12–14 → ai.engineer/nyc/2026

Time to take the conversations off stage. We're wrapping up Day 1 of #AIEParis with our Speaker Dinner hosted with Notion, bringing today's speakers together around the table for good food, great company and conversations beyond the stage. See you around the table. 🥂 @NotionHQ

AI Engineer Shanghai | Nov 5-6, 2026 Join @amrcn_werewolf (@Microsoft), @ryolu_ (ex-@cursor_ai), @chlassner (@theworldlabs), @NancyZWang (@1Password) & @liu8in (@HeyGen). Open models, AI infra, Physical AI. Technical talks + demos. Tickets: ai.engineer/shanghai/2026

I'm next on stage @ @aiDotEngineer lfg