StarkWare Ecosystem Lead Departs to Focus on AI Safety
TREE NEWS reports: Abdelhamid Bakhta, formerly the ecosystem lead at StarkWare, has announced his departure from the company. In a public statement, Bakhta said he will dedicate his next chapter to AI safety, citing a growing conviction that the field represents one of humanity’s most pressing challenges. His exit marks a notable shift for a key figure in the StarkNet ecosystem, which has been at the forefront of zero-knowledge scaling solutions for Ethereum.
From ZK-Rollups to AI Safety
Bakhta’s move underscores a broader trend of crypto talent migrating toward artificial intelligence, particularly in areas where cryptographic techniques can address trust and verification problems. StarkWare, known for its STARK-based validity proofs, has long championed the idea that zero-knowledge cryptography can enhance privacy and scalability. Bakhta’s new focus suggests that similar cryptographic primitives could be applied to AI safety—for instance, in verifying model behavior, ensuring data provenance, or creating auditable AI systems.
During his tenure, Bakhta played a role in expanding StarkNet’s developer ecosystem, which has attracted projects in DeFi, gaming, and identity. His departure raises questions about the continuity of those efforts, though StarkWare has a deep bench of engineers and ecosystem contributors. The company has not announced a direct replacement.
AI Safety as the Next Frontier
Bakhta is not alone in pivoting to AI safety. Several crypto researchers and founders have recently argued that blockchain-based verification and decentralized compute can help mitigate risks from advanced AI. Projects focused on zero-knowledge machine learning (zkML) and decentralized AI networks are gaining traction, promising ways to prove that models ran correctly without revealing sensitive data. Bakhta’s expertise in STARKs could be particularly relevant to these efforts, as STARKs offer transparent, post-quantum secure proofs.
However, the intersection of AI safety and crypto remains nascent. Critics argue that many AI safety challenges—such as alignment and interpretability—are not easily solved by cryptography alone. Bakhta’s success will depend on whether he can bridge the gap between cryptographic guarantees and the messy reality of modern machine learning.
Forward-Looking Perspective
As AI capabilities advance rapidly, the demand for robust safety measures is likely to grow. Bakhta’s move signals that crypto-native talent sees an opportunity to contribute. Whether through decentralized compute networks, verifiable inference, or on-chain AI agents, the crypto industry is increasingly positioning itself as a source of infrastructure for trustworthy AI. StarkWare’s loss may be AI safety’s gain—and a sign that the lines between crypto and AI are blurring.
