Ethereum Foundation Hires Erigon's Giulio Rebuffo to Push AI-Written Code

The Ethereum Foundation has hired Giulio Rebuffo, former CTO of the Erigon client and one of the protocol's youngest-ever core developers, for a new role focused on AI-generated code, testing and verification, according to his own announcement of the move.

The hire lands at a strained moment for the organization. Rebuffo joins a foundation that has lost at least eight high-profile staff this year and cut 20% of its workforce, with a further 40% budget reduction confirmed by co-founder Vitalik Buterin in June.

Who is Giulio Rebuffo?

His résumé reads unusually for a foundation appointment. In an October 2025 interview, Rebuffo recounted joining TurboGeth — the project that later became Erigon — while still in high school. At 17, he became one of Ethereum's youngest core developers.

He served as principal engineer on Erigon during the Merge, Ethereum's 2022 transition to proof-of-stake, and took over as the client's CTO by 22. His work centered on prototyping features and extracting more speed from the protocol without changing it through hard forks.

The skill he names as most important is adversarial thinking. He assumes any code he writes will be attacked. That instinct now feeds directly into his mandate: making machine-written code safe enough to run on a chain that settles hundreds of billions of dollars in value.

What does the hire change for the foundation?

Rebuffo will work across technical teams and management, per his post, placing him close to the client developers who maintain Ethereum's execution software. His brief covers AI tooling, testing and formal verification — the practice of mathematically proving code behaves as intended.

The appointment maps onto an argument Buterin has pressed for months. In March, Buterin called an experimental AI-built client "quite an impressive experiment" after developer Jiayao Qi assembled ETH2030, a 702,000-line Go client targeting Ethereum's draft 2030 roadmap — built in roughly six days using Claude Code for about $5,750.

Buterin said teams should "take half the gains from AI in speed, and half the gains in security." He has pushed for more test cases, more formal verification and more independent implementations of the same software.

In a May blog post on formal verification, Buterin went further. Done well, he wrote, it could produce code that is "far more secure than the way programming has been done before." He framed AI-assisted verification explicitly as a defense against AI-assisted attacks on crypto infrastructure.

Eight exits, a lawsuit and a shrinking budget

The foundation Rebuffo is joining bears little resemblance to last year's roster.

• Researchers Julian Ma and Carl Beek resigned in May.
• Tomasz Stańczak stepped down as co-executive director in February, after less than a year in the job.
• Josh Stark and Protocol Guild organizer Trent Van Epps announced their exits the same day in April.
• The three leads of the influential Protocol cluster — Barnabé Monnot, Tim Beiko and Alex Stokes — moved on or took sabbatical, prompting the foundation to install Will Corcoran, Kev Wedderburn and Fredrik as new cluster coordinators.
• Geth creator Péter Szilágyi had left the previous year, after roughly a decade.

A former employee also sued the foundation and Buterin over a firing earlier in November.

Observers hoping for high-profile replacements for departed executives may have to wait. The June cuts — a fifth of staff and 40% of the budget — have already produced think pieces questioning whether the foundation can still deliver on the blockchain's technical roadmap.

One hire, not a rebuild

For now, the foundation has hired one engineer with deep client experience and pointed him at AI tooling, testing and verification. That is a bet on a specific thesis: that the next generation of Ethereum's software will be written — and checked — substantially by machines.

Rebuffo's adversarial instincts will get an early test. Whether AI-written clients like ETH2030 can survive hostile scrutiny, and whether the slimmed-down foundation can verify them fast enough to matter, will become clear as Ethereum's 2030 roadmap moves from draft to implementation.