AI Agent Olympics Hackathon Underway at Milan AI Week

Today, selected builders are on-site at Milan AI Week 🇮🇹 for the AI Agent Olympics Hackathon from lablab.ai. Teams are developing autonomous AI agents and enterprise AI applications in real time at Fiera Milano, powered by Vultr infrastructure. The future of agentic AI is being built this week. https://lnkd.in/dSWsXJR5

We're in! 🪶 Excited to join the AI Agent Olympics Hackathon by lablab.ai and challenge builders to go deep, not wide. The era of general-purpose assistants is over, we want to see agents that own a single domain completely. Legal research. Medical triage. Code review. Logistics. Content moderation. Pick your lane and build something production-shaped on top of 30,000+ open-source models! One OpenAI-compatible API, zero GPU management. Part of our team will be on the ground in Milan 🇮🇹! Come find us, talk shop, and show us what you're building. Neutral infrastructure. Open models. Specialist agents. Let's go.

I'm in Milan for AI WEEK as Featherless AI is sponsoring the lablab.ai AI Agents Olympics Hackathon! Fresh off our $20M Series A co-led by AMD Ventures and Airbus Ventures, and now putting it to work where it matters: in the hands of builders. In town all week. If you're in Milan, let's grab a coffee ☕, would love to meet founders, builders, and anyone curious about where open-source AI is headed.

Hands-on work often clarifies where AI is headed faster than any roadmap. The Agentic Hackathon hosted by Google Public Sector, SZNS Solutions and Carahsoft gives Government teams the space to build, test and refine autonomous AI agents that operate in real environments, not just theory. Across two days at the Google Office in DC, attendees can also earn up to 6.6 CPE credits. For teams exploring agentic workflows, it’s a direct way to see what’s possible when the right tools and approach come together. Details and registration: https://ow.ly/aNSs30sV3Yj

Hands-on work often clarifies where AI is headed faster than any roadmap. The Agentic Hackathon hosted by Google Public Sector, SZNS Solutions and Carahsoft gives Government teams the space to build, test and refine autonomous AI agents that operate in real environments, not just theory. Across two days at the Google Office in DC, attendees can also earn up to 6.6 CPE credits. For teams exploring agentic workflows, it’s a direct way to see what’s possible when the right tools and approach come together. Details and registration: https://ow.ly/1L0b30sVGEq

Winner at DT Global hackathon 2026 is great. But what matters more is what it proves.. In 3 days, we built Nexus Intelligence — a fully functional AI application..not a concept, not a demo. Most organizations are still “using AI.” Very few like DigitalTolk are building AI into how they operate. 𝗪𝗵𝗮𝘁 𝘄𝗲 𝗳𝗼𝗰𝘂𝘀𝗲𝗱 𝗼𝗻: We didn’t start with models. We started with problems inside our system. → fragmented internal knowledge → repetitive operational workflows → dependency on people for context So we built a layer that sits on top of everything. Nexus Intelligence works in two clear modes: • Guide Mode — understands, reasons, assists • Action Mode — debugs, fixes, executes, automates, drives outcomes Powered by RAG, it grounds every response in real internal data — making AI usable in actual workflows, not just conversations. 𝗪𝗵𝘆 𝘁𝗵𝗶𝘀 𝗺𝗮𝘁𝘁𝗲𝗿𝘀 (beyond a hackathon): AI will not create advantage by itself. Every company will have access to the same models. The advantage will come from: → how well your AI understands your internal context → how deeply it is embedded into your workflows → how fast your teams can move from idea → execution That’s where internal tools + AI systems become a strategic differentiator. 𝗪𝗵𝗮𝘁 𝘀𝘁𝗼𝗼𝗱 𝗼𝘂𝘁 𝘁𝗼 𝗺𝗲 𝗽𝗲𝗿𝘀𝗼𝗻𝗮𝗹𝗹𝘆 Speed is no longer the differentiator. Clarity + Context + High trust + Collaboration is. Huge shoutout to my teammates Shivesh Tripathi Mourya Tanuku — this was pure ownership and execution under commitment. And to DigitalTolk , Virpal Singh Günter Grodotzki Samuel L. for creating a space where building like this is possible. This is just the beginning. Internal AI systems are going to redefine how engineering teams operate. . . . #Engineering #EngineeringIntelligence #EngineeringExcellence #AI #InternalSystems #RAG #Hackathon

An hour into this year's AI Hackathon at DG, and the place is buzzing. Great to hear so many lively discussions on how best to tackle a real-life problem using AI. The five student teams are already impressing the judges with their approaches. It's going to be a close contest.

Speed to prototype was never the only bottleneck. Anyone who's run a hackathon knows it. Twenty-four hours to a working demo, and then months to ship — if it shipped at all. The prototype was never the only hard part. Ownership, observability, cost at scale, team capability to maintain what got built — that's where things stalled. AI compresses the prototype phase by orders of magnitude. The rest of the bottlenecks are still there. You just arrive at them faster now, and more often. The partners worth working with are the ones who've solved the hard part, not just the fast part. #GenerativeAI #AgenticAI #AWSPartner #EnterpriseAI #CommerceArchitects

This is a deja vu moment: new tool/process comes along speeding up initial development. Everyone gets excited, demos delight, leadership makes grand plans for a massive roadmap. And then the same bottlenecks show up. AI is a bigger accelerator than the previous leaps, but the same pattern is here. Shipping high-value, scalable solutions to production has been in our DNA from the start. That focus is more critical than ever right now.

The part AI has changed: you can have a working prototype in hours. The part AI hasn't changed: All of the systems have to exist around it. Prompt behavior drifts in production and you have no visibility into why. Latency looks fine in a demo and falls apart under real load. The data pipeline that worked against a clean dataset meets messy reality. Costs scale in ways nobody modeled because nobody had to yet. Fast prototyping is only the first step. How you manage production systems,. up-time, scaling, and resilience still takes experience and expertise.