Manos Koukoumidis

Train custom AI with Oumi and deploy in Amazon Web Services (AWS) Bedrock - end-to-end in just hours! It has never been easier to build a model that is optimized to your use case to beat the performance (quality, latency, cost) of off the shelf models, and then deploy to production. An end-to-end solution for custom AI is only possible with close integration between software and hardware. We’re excited to share our joint work with Amazon Web Services (AWS) showing how to synthesize training data and fine-tune an LLM using Oumi on Amazon EC2, store artifacts in Amazon S3, and deploy to Amazon Bedrock. Very excited for this partnership with the amazing Amazon Web Services (AWS) team. Deb Das, Bashir Mohammed, PhD, Bala Krishnamoorthy, Gregory Fina, Deap Ubhi, and Pronoy Chopra, thank you for the outstanding collaboration! Link below.

Weekend news: OpenClaw just got clawed (acquired) by OpenAI! : 🦀 Meta was trying to get Peter Steinberger of OpenClaw to join as well - personal agent competition heating up. 🔥 @ChandanLodha predicts the following for the future of OpenClaw - what do you think? OpenClaw CHANGELOG: v1.0 — Open source launch v1.1 — Sam calls founder "a genius" v2.0 — Joins "foundation" v2.1 — Foundation requires API key v3.0 — API key requires enterprise contract v4.0 — "We've always been committed to openness" v5.0 — import openclaw → ModuleNotFoundError

Just had a PR merged to @oumi_ai! Multi-turn data synthesis support is now in the codebase. This enables programmatic generation of full conversational dialogues. Here's what's coming and why researchers will care 🧵 github.com/oumi-ai/oumi/p…

Want a prediction for the Super Bowl or your future based on your AI strategy? Ask ChatGPT. 🏈 🏈 🏈 ❓ Respond in a single sentence. Who will win the Superbowl 2026? ❗ The Seattle Seahawks are widely favored by odds and expert predictions to win Super Bowl LX (2026) over the New England Patriots. 🤖 🤖 🤖 ❓ Respond in a single sentence. Which enterprise will win, the one only prompting closed off the shelf models or the one only fine-tuning open models? ❗ The enterprise that only fine-tunes open models will win, because ownership of the model stack creates compounding advantages in cost, performance, and strategic independence that pure prompting can’t match. You heard ChatGPT... (Feb 8th 2026, 3:21pm) The one thing that ChatGPT doesn't know "yet" is how easy it is to finetune models using Oumi. What are you waiting for? 🚀

Stop renting intelligence built for everyone else, build the intelligent that is right for your enterprise. The enterprises that will win, will be the ones that lean in to build their own intelligence to establish and strengthen their differentiation. The ones that are curating the same prompt for the same off the shelf model that everyone else is using will end up on the wrong side of history. Building specialized intelligence was a privilege of the few. But not any more... With the Oumi + Lambda partnership, enterprises are getting for the first time the ability to build specialized models 100x faster that before. Yes 100x, not just 10x.

Enterprises are constrained by large, off-the-shelf AI models that are costly, slow, and hard to control, while custom models promise better performance but traditionally take months to build. Today, Oumi and Lambda announced a partnership that changes this, enabling enterprises to build and deploy high-performance custom AI models in hours, not months. The result: up to 70% cost reduction, faster iteration, and full control over data, privacy, and infrastructure. Read more here: blog.oumi.ai/p/lambda-and-o…

OpenThoughts-Agent Blog: openthoughts.ai/blog/agent Why Oumi Blog: blog.oumi.ai/p/if-ai-isnt-t…

OpenThoughts-Agent just launched with open code, data and weights achieving SOTA across open weight models of its size for both Terminal Bench and SWE-Bench! 🚀 It is a testament to what strong open source AI collaborations can deliver. We at Oumi are proud to be part of this effort and excited for what it means. Such achievements are great news for enterprises, sciences and humanity overall. They are helping put AI on on a better path – the reason why we founded Oumi. Truly exceptional work by so many people, universities and companies on OpenThoughts-Agent. 👏

It’s here. AI Agents Assemble, the biggest WeMakeDevs hackathon so far. $20,000 (₹18 lakh) in cash prizes 💰 Winners get job interview opportunities. Everyone who participates gets Google Summer of Code mentorship ☀️ First place alone wins $10,000 cash 🥇 Powered by some of the biggest names in tech: @cline, @kestra_io, @vercel, @Oumi_PBC, @coderabbitai and @togethercompute! You’ll learn a lot, build with real AI tools and meet incredible people from the community. All skill levels are welcome. If you’ve ever wanted to level up, this is the perfect place to do it. 👉 Register: wemakedevs.org/hackathons/ass… 🗓️ December 8 - 14, 2025 Tag someone you want to build with or someone who should take part 👇

Airbnb CEO Brian Chesky: “We’re relying a lot on Alibaba’s Qwen model. It’s very good. It’s also fast and cheap... We use OpenAI’s latest models, but we typically don’t use them that much in production because there are faster and cheaper models.” The valley is built on Qwen?

Yesterday, many world-class AI researchers, engineers and former colleagues were impacted by layoffs at Meta Facebook . If you’re one of them — or know someone who is — please reach out or share this. I’d be happy to connect you with friends across the AI ecosystem or explore opportunities together. If you’re looking to join a truly open AI lab, we’re hiring at Oumi. We’re building the AI future worth building — one that’s transparent, collaborative, and community-driven. 🚀

Massive thanks to @FastCompany for recognizing Oumi in the “Next Big Things in Tech 2025” 💬 As @harrymccracken from Fast Company puts it: “At the moment, much of AI’s progress feels like it’s driven by fierce competition among a handful of giants intent on controlling their entire technological stacks. Oumi has a different vision: It’s creating a completely open platform for foundational models that includes data curation, training, and inference. Proud to be receiving this award alongside stellar recipients like @cartesia , @runwayml , @turingcom and others. 🏆 🏆 🏆

The blog by Stefan Webb is worth reading: lnkd.in/guPjtqhq

Small fine-tuned models are (most often) all you need, and the evidence is overwhelming. It may be a bit too early for 2026 predictions, but 2026 will be the year of specialized SLMs. I am calling it.

New blog post: Hours, Not Months – The Custom AI Era is Now: open.substack.com/pub/oumiai/p/h… Oumi website: oumi.ai

When we started Oumi, we set out to build a better future for AI – grounded in the benefits of open source. Eight months ago, we open-sourced the most comprehensive repo for training foundation models end to end. Today, we’re building on that by enabling every developer and enterprise to train custom models in hours, not months ⏱️ – a foundational shift in how teams build, ship, and own AI. I shared why custom models are the future – and why now is the moment – in our new blog post (see first comment for link). 📖 Ready to jump to the new era? Sign up for early access and join our limited design partner program. See more at our refreshed website. Proud of the world-class Oumi team: builders and operators who are shaping what comes next. 🚀

🏁 The winners will be announced on October 22, 2025. For more information about the CODiE Awards and to view all finalists, visit the CODiE website (link at first comment). Learn more about Oumi and why not give us a ⭐️ over at GitHub (link at first comment)? Oumi GitHub: github.com/oumi-ai/oumi CODiE awards: codieawards.com/finalists

🎉 Thrilled to share that @Oumi is a finalist for the @CodieAwards in the category “Best Open-Source Development Tool”! 🎉 #CODiEAwards 🤓 The CODiE Awards are the premier peer-recognized program celebrating innovation and excellence in technology. Each year, industry experts evaluate products for their innovation, impact, and overall value. 🤝 We’re grateful to CODiE for recognizing Oumi and our mission, to help lower the barrier-to-entry for foundation model development. Our vision is to democratize the building of robust and cost-effective GenAI applications.

Today, The White House released America’s AI Action Plan—a pivotal step in shaping the future of AI leadership. To stay ahead in the global AI race, the plan emphasizes the need to “Encourage Open-Source and Open-Weight AI.” While companies like OpenAI and Anthropic continue to push for centralized, closed-source control of frontier models, it's refreshing to see the U.S. government cutting through their self-serving 💩  and recognizing the value of openness and decentralization in AI development. Thank you @MichaelKratsios and team! This is a strong signal: open innovation is national strategy. Let’s double down. Let’s build AI in the open—together. 🧑‍🤝‍🧑 🚀