Ines Montani 〰️

inesmontani.bsky.social

12 October 2026 13:24

Ever since we started woking on bringing our ideas into the cloud, we knew we wanted a hybrid architecture that lets users plug in their own compute. A bit like self-hosted runners in a CI system. For our new tool Ellf, we host the app and you plug in your own data-private cluster 🧵

Ines Montani 〰️

inesmontani.bsky.social

05 August 2026 07:59

Attention Berlin people: On August 10 (next Monday), I'm co-hosting a meetup with Hugo and Jodie @t-redactyl.bsky.social at @jetbrains.com 🚀 Featuring: agents, agentic NLP, coding assistants and lightning talks. See you there! 📆 Mon, Aug 10, 6–9 pm 📍 JetBrains, Berlin ✅ RSVP: luma.com/agmd8ie4

Ines Montani 〰️

inesmontani.bsky.social

27 July 2026 14:31

Ellf gives the agent skills + MCP servers + agents with our distilled expertise from 10+ years of solving NLP problems: a virtual NLP engineer. It's backed by a compute platform plus your own private cluster, local or cloud, and uses Prodigy for annotation – by humans and agents. beta.ellf.ai

Ines Montani 〰️

inesmontani.bsky.social

27 July 2026 14:31

Software 1.0: code → compiler → program. Coding agents are already pretty good at this. Software 2.0: code + data → training → model. Requires running lots of stuff: annotation, preprocessing, evaluation, analysis, ... So agentic software 2.0 needs a new approach.

Ines Montani 〰️

inesmontani.bsky.social

27 July 2026 14:31

If building modular systems is so obviously better than "LLM maximalism", why don't we all just build classifiers? Because it's lots of steps, annoying, operationally complex, and needs data and expertise. So we asked: could coding assistants and agents make this as efficient as writing code?