←
DataRobot11,690 posts
DataRobot avatar

DataRobot

@DataRobot

the only agent workforce platform built for outcomes — not endless pilots.

◷ Joined June 2012▧ 3,245 Photos and videos
Supply Chain Now avatar
⟳ DataRobot retweeted
Supply Chain Now@_supplychainnow·Sep 18

AI can flag a supply chain risk. But what happens next? The real value comes when AI can help teams act on the problem. @ScottwLuton and Kim Reuter talk with Vika Smilansky and Marc Amarillas of @DataRobot about using AI agents to tackle supplier risk, tariffs and operational disruptions, including a $2.5B risk addressed in 72 hours. 🎥 Watch on demand: hubs.ly/Q04xP9vy0

1:01
DataRobot avatar
DataRobot@DataRobot·Sep 14

Most agents get tested once, by whoever built them, on prompts that work. No signal on a malicious payload or six turns of pushback. Agent Assist now red-teams agents before deploy: multi-turn attacks, fixes you approve, a readiness verdict for the PR. datarobot.com/blog/adversari…

Supply Chain Now avatar
⟳ DataRobot retweeted
Supply Chain Now@_supplychainnow·Sep 1

Most AI pilots never make it past the test phase. What happens when you actually put AI to work? Launched today: @ScottwLuton, Kim Reuter, Vika Smilansky and Marc Amarillas of @DataRobot break down what it takes to move agentic AI from pilot to production, including a real $2.5B supply chain risk resolved in 72 hours. Watch the new on demand webinar: hubs.ly/Q04w6B_W0

Post attachment
mojesko avatar
⟳ DataRobot retweeted
mojesko@mojeskoqq·Aug 3

Elvin Aghammadzada, DataRobot, mapped where a context window stops working: up to 40% full it's the smart zone, past that it's the dumb zone here's how to shrink an agent's context, from scratch: step 1 → measure your own edge, 2026 measurements put the start of degradation at 25% of the window, not near the end - a window is not a tank, it's a runway step 2 → count what's burned before the user says a word, an agent wired to 15 MCP servers spends over 100,000 tokens on tool definitions alone step 3 → load on demand, a skill's front matter is under 100 tokens and the body only arrives when it's needed, roughly 10x cheaper in context than plain MCP step 4 → put teachability on the buying checklist next to security and compliance, skills already ship natively in 26+ platforms step 5 → prune the junk, model-written skills perform worse than human-written ones and there is still no verification layer for them on August 1-2 OpenAI showed Astra: ten problems open for decades, cracked by coordinating multiple agents over long runs, with about $2,000 of tokens for all ten most people bolt on one more MCP server - he switches off everything the current step doesn't need bookmark & watch - this 26-min talk, then read the full context rot breakdown below ↓

2:30

mojesko @mojeskoqq Jul 13 Article The More Context You Give, The Dumber The Model Gets A quiet Tuesday. The agent has been live for two weeks, running on a 1M-token window, fed 320,000 tokens of transcripts, notes and style guide on every call. And it keeps citing the wrong quarter. The

DataRobot avatar
DataRobot@DataRobot·Aug 7

Our own Jyothi Nookula breaks down why an agent failing in production isn't a platform problem or a developer problem. It's a secret third thing😎

3:03
DataRobot avatar
DataRobot@DataRobot·Aug 5

We've been asking our devs: who's actually responsible when an agent fails in production? The developer, the platform, or the org? One dev's perspective:

1:05
DataRobot avatar
DataRobot@DataRobot·Aug 4

Not to name names 🫣 but who's actually responsible when an agent fails in production? The developer, the platform, or the org? One dev's take:

1:16
DataRobot avatar
DataRobot@DataRobot·Aug 3

Dev to dev: your agent isn't broken. Your testing process is. The teams making it to production are the ones systematically catching errors, turning them into tests, and iterating before they become production fires. One DataRobot dev built a framework for exactly that:

1:47
DataRobot avatar
DataRobot@DataRobot·Aug 3

Trustworthy AI should give similar answers to similar questions. If your agent's responses vary wildly, you've got a stability issue. Hear about how we've been tackling the stochastic nature of language models head-on:

1:47