Hiring Kaggle Masters is $$$.
Hiring NEO… isn’t.
If this scales, it’s a nightmare for data science job boards. https://t.co/hcfiElwMzT
1/ Introducing @withneo — the groundbreaking autonomous ML engineer.
A complete solution, not just a copilot or autocomplete.
It handles the entire pipeline:
→ Prepares data
→ Trains models
→ Debugs pipelines
→ Submits results
Truly end-to-end.
OpenAI created MLE-Bench for ICLR '25, inspired by 75 real Kaggle competitions.
The test was straightforward:
An agent demonstrating effective ML engineering.
3/ Results:
Microsoft RD-Agent → 22.4%
NEO (6 people, $500k) → 34.2%
Same benchmark.
Same rules.
David didn’t just win.
He embarrassed Goliath.
4/ Why this matters:
→ ML engineers: less grunt work, more research
→ Startups: Kaggle-level talent without $$$
→ Industry: first time an agent beat a trillion-dollar lab at real ML engineering
5/ Think weeks of ML work done in hours.
Agents handle the grind.
Humans focus on ideas.
This is the Devin/Cursor moment for ML engineers.
6/ Watch NEO in action
Sign up for early access:
https://t.co/kAq5SBU4dR
I hope you've found this thread helpful.
Follow me @abxxai for more.
Like & Repost the post to help others:
