BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//hackathonmaxxing//EN
CALSCALE:GREGORIAN
X-WR-CALNAME:Hackathons in New York
BEGIN:VEVENT
UID:luma:evt-PxLMfa23NSKhTr5@hackathonmaxxing
DTSTAMP:20261122T162727Z
DTSTART:20261203T230000Z
DTEND:20261204T013000Z
SUMMARY:The Real Cost of AI: Lessons from Teams Running It at Scale
URL:https://lu.ma/21l5tw57?utm_source=hackathonmaxxing&utm_medium=referral&utm_campaign=new-york
LOCATION:New York\, NY
DESCRIPTION:This event will be held in Hebrew\n\nAn evening for founders and technology leaders building with AI on AWS\n\nThursday\, December 3\, 20266:00 PM to 9:00 PM860 Broadway\, New York City\n\nJoin Directeam and Bereisheet for an evening of practical talks from engineers running AI at real scale. No hype\, no product pitches\, just what actually works.\n\nWhat you’ll learn\n\nHow a video intelligence service indexes 8 million minutes a day at about 40x lower cost than the naive approach\n\nHow to measure R&D productivity in the AI era\, beyond counting tokens\n\nHow to build feedback loops that make agents better and cheaper over time\n\nWhere AI spend leaks: agent economics\, cost attribution\, and the costs that only show up at scale\n\nA live panel with all three speakers: bring your hardest AI cost question\n\nSessions\n\nThe Economics of Video Intelligence: Understanding 8 Million Minutes a Day Without Breaking the Budget | Yaron Gurovich\, VimeoVideo is the most expensive modality to understand with LLMs. Yaron shows how breaking the problem into small\, well-scoped steps lets cheaper models do most of the work\, why the right pipeline depends on the model\, and how benchmarks make it safe to swap models.\n\nScaling AI Adoption in Engineering | Lior RapaportA real case study of AI adoption inside a large engineering org: measuring output\, managing agent permissions so data doesn’t leak\, building a “second brain” as a single source of truth\, and what happens when autonomous agents run up a six-figure bill.\n\nHow We Make Our Agents Better and Cheaper in Production | Ori Stoliar\, CTO\, DirecteamHow Directeam built its agent feedback loop: which flows run\, how they’re measured\, and how answers keep improving. Plus the cost side: optimizing spend during feedback loops (including the Price Reversal Phenomenon) and why cost attribution matters.\n\nPanel: The Real Cost of AI in Production | Ori\, Yaron & LiorThree engineers\, three very different AI workloads\, one shared problem. A candid conversation about where https://lu.ma/21l5tw57?utm_source=hackathonmaxxing&utm_medium=referral&utm_campaign=new-york
END:VEVENT
END:VCALENDAR
