Mete Polat

founder @VartoAI. Prev. formal verification at Apple, Cambridge, TUM.

[Mete Polat @MetePolat]

This was a side gig for us. More on agentic verification of real-world software soon.

[Varto @VartoAI]

Varto and GPT-5.5 proved the complete PutnamBench math AI benchmark, formally verified in Isabelle
https://t.co/t82mngzOin

[Mete Polat @MetePolat]

@VartoAI Proving competition math formally now costs just $8 and 16 minutes on average.