Inside the Loop with Mercor: RL Environments, Benchmarks, and Evals

Inside the Loop: Evals → Real-World Data → RL Environments → Better Models A Meet & Greet with Engineering Leadership at Mercor NY 📅 Tuesday, October 20 · 7:00–9:00 PM The way frontier models get better has changed. Scaling pre-training used to be the main driver of progress. Now a lot of the gains come from post-training: rigorous evals, high-quality real-world data, and RL environments that test models on the work people do. Mercor sits inside that loop. We work with leading AI labs, so we get an early look at where models are falling short and what the next generation needs to learn. Join engineering leaders from Mercor and Deeptune for a candid fireside chat on the flywheel we’re building and what it shows about where AI is going. What we’ll cover How model improvement has changed: why the frontier has moved from “more compute” to “better signal” Why post-training is critical: how evals and real-world data turn into measurable gains in model capability Building RL environments…