This interview is with Marouen Zelleg, Co-Founder at Crestal.

Marouen Zelleg, Co-Founder, Crestal

Marouen, for those who may be unfamiliar with your work, can you provide a brief overview of your background and expertise in the world of blockchain, crypto, Web3, and AI?

My career has been driven by a passion for solving complex technology challenges and bridging innovation with real-world impact. Over the past decade, I’ve worked at the intersection of cutting-edge technologies, starting with enterprise IT and progressing to blockchain, crypto, and Web3. My early experience revolved around helping businesses modernize by decommissioning legacy systems and launching digital products across Asia. This foundation set the stage for deeper work in blockchain and Web3, where I held leadership roles at ConsenSys and Polygon Labs, driving strategic partnerships and scaling adoption of Ethereum-based solutions. Founding Crestal was a natural evolution of my journey—combining blockchain, AI, and infrastructure to create scalable, decentralized ecosystems tailored for real-world use cases. My expertise lies in designing systems that empower businesses to navigate and thrive in the decentralized future.

What pivotal moments or experiences led you to focus on these cutting-edge technologies and ultimately shaped your career trajectory?

One of the most defining moments in my career was during my time at OutSystems, where I witnessed firsthand the challenges businesses faced in modernizing their operations amidst IT talent shortages and growing complexities in legacy systems. This experience pushed me to explore solutions that combined innovation with scalability. My transition to blockchain was catalyzed at ConsenSys, where I saw the transformative potential of decentralized technologies to reimagine trust and efficiency in industries. At Polygon Labs, another pivotal moment came when we worked on scaling Ethereum with zero-knowledge technology, a process that highlighted how combining advanced cryptography with Web3 principles could unlock massive scalability and accessibility. These experiences instilled in me a belief that blockchain and AI, when applied thoughtfully, could solve some of the most pressing technological and societal challenges. Founding Crestal was the culmination of this vision—to design systems that not only leverage these technologies but also make them accessible and actionable for businesses of all sizes.

From your experience at Crestal, how can the innovative combination of onchain, offchain, and aggregated data be applied to provide a more holistic view of other industries beyond blockchain infrastructure?

At Crestal, we’ve seen how integrating on-chain, off-chain, and aggregated data creates a more comprehensive understanding of performance, trust, and decision-making in decentralized systems. This approach is not limited to blockchain infrastructure—it has transformative applications across industries. For example, in supply chain management, on-chain data can ensure transparency and traceability of goods, while off-chain data like supplier performance or environmental factors adds critical context. Aggregating these data sets allows companies to predict disruptions, optimize logistics, and comply with sustainability standards. Similarly, in healthcare, on-chain data can securely manage patient records, while off-chain data from IoT devices and aggregated insights from clinical studies provide a 360-degree view of patient health. This can lead to more precise, proactive care. The key lesson is that combining these data layers provides a holistic perspective that helps industries transition from reactive operations to predictive and adaptive systems, improving outcomes across the board.

You mentioned the development of an Architecture Design Assistant AI model. Can you elaborate on an instance where this model provided unexpectedly valuable insights or guidance during a project?

One instance that stands out is when we were designing a decentralized infrastructure solution for a client with highly-specific scalability requirements. The Architecture Design Assistant AI model analyzed their workload patterns, predicted potential bottlenecks, and recommended a hybrid infrastructure configuration—something we hadn’t initially considered. It suggested allocating certain resource-intensive operations to off-chain environments while keeping critical verification tasks on-chain. This not only optimized performance but also reduced operational costs by over 30%. The AI model also flagged underutilized resources in real-time, enabling the team to fine-tune the deployment dynamically. What was unexpected was the level of granularity and actionable insights the model provided, which significantly accelerated the decision-making process and enhanced the project’s efficiency. This experience reinforced the value of combining AI-driven insights with human expertise to design highly-tailored, scalable solutions.

You learned a valuable lesson about AI not being a one-size-fits-all solution. What practical steps can businesses take to ensure their AI implementations are tailored to the specific nuances of the Web3 landscape?

One critical step is to invest in domain-specific data. In the Web3 landscape, every use case—whether it’s decentralized finance, gaming, or infrastructure—has unique patterns and challenges. Training AI models on generalized data often leads to inaccurate insights, so businesses must curate datasets that reflect the specific nuances of their application.

Second, adopt a modular AI approach. Instead of deploying monolithic AI systems, design solutions that can be adjusted or augmented based on evolving needs. For example, at Crestal, we developed our AI tools to allow flexibility, such as integrating additional data sources like on-chain performance metrics or off-chain developer behavior.

Finally, maintain human oversight and feedback loops. Web3 ecosystems are inherently dynamic, and AI must evolve with them. Regular audits, iterative training, and user feedback are vital to ensure the AI remains aligned with the organization’s goals and delivers actionable, trustworthy insights.

Data security is paramount. Beyond end-to-end encryption and role-based access controls, what emerging technologies or strategies are you most excited about for safeguarding customer data in the decentralized world?

One of the most promising emerging strategies is the use of zero-knowledge proofs (ZKPs). These allow data to be verified without revealing the underlying information, ensuring privacy while maintaining trust. At Crestal, we see ZKPs as a game-changer for secure transactions and identity verification in decentralized systems. They enable users to prove eligibility or ownership without sharing sensitive data, reducing exposure to breaches. Another exciting development is secure multi-party computation (SMPC), which allows multiple entities to jointly process data without revealing their inputs. This is particularly useful in decentralized environments where collaboration without trust is essential. Additionally, decentralized identity (DID) frameworks are making strides in empowering users to control their own data. By leveraging blockchain and cryptographic keys, users can share only the data necessary for a specific interaction, minimizing exposure. Combining these technologies creates a robust and user-centric approach to data security that aligns with the ethos of decentralization.

How can AI, beyond just real-time monitoring, be leveraged to proactively identify and address potential security vulnerabilities before they are exploited in a blockchain network?

AI can be leveraged proactively by using predictive analytics and anomaly detection. By analyzing historical on-chain and off-chain data, AI can identify patterns that typically precede vulnerabilities, such as irregular transaction behavior, network latency spikes, or unexpected resource consumption. This allows for early intervention before these indicators escalate into full-blown threats. For example, at Crestal, we use AI to continuously simulate potential attack vectors in a sandboxed environment. This "red team" approach enables the system to identify weak points in smart contracts or infrastructure setups before they are deployed. Another effective method is leveraging AI for code review, particularly for smart contracts, where vulnerabilities like reentrancy attacks can be detected before they are exploited. AI can also improve the response by automating mitigation strategies. For instance, if an anomaly is detected, the system could automatically limit permissions or isolate affected nodes, reducing the attack surface while alerting human operators. By transitioning from reactive to proactive security measures, AI becomes a powerful tool for safeguarding decentralized networks.

You've mentioned how AI-driven customer insights shaped Crestal's product roadmap. What advice would you give to startups struggling to effectively gather and utilize customer data to inform their product strategy in the rapidly evolving Web3 space?

Startups need to start with a clear understanding of their user personas and behaviors. In the Web3 space, where users often interact pseudonymously, focusing on on-chain data such as wallet activity, transaction history, and protocol interactions can provide valuable insights. Complement this with off-chain feedback from user communities, like forums and social media, to understand pain points and expectations.

One strategy we've employed at Crestal is leveraging AI to cluster and analyze user behavior patterns. For example, identifying how developers interact with infrastructure tools helped us prioritize features that optimized deployment speed and performance monitoring. Startups can use similar approaches by creating feedback loops where data informs product decisions, and product updates in turn generate new insights.

Additionally, fostering direct engagement with your community through platforms like Discord or Twitter can uncover qualitative insights that AI alone might miss. Marrying these approaches—quantitative analysis via AI and qualitative input from users—enables a more holistic understanding of customer needs and ensures your product remains relevant and valuable.

Looking ahead, what are some significant opportunities you see for the convergence of blockchain, crypto, Web3, and AI, and what impact do you believe these technologies will have on the venture capital landscape in the coming years?

The convergence of blockchain, Web3, and AI presents a wealth of opportunities, particularly in creating decentralized autonomous systems that are smarter, more adaptive, and capable of self-governance. AI can enhance blockchain scalability through predictive resource allocation, optimize smart-contract execution, and improve user experience by personalizing decentralized applications (dApps). One significant opportunity lies in decentralized finance (DeFi), where AI can provide real-time risk assessment, detect fraudulent activity, and dynamically adjust lending protocols based on market conditions. Another area is supply chain management, where combining AI’s analytical capabilities with blockchain’s transparency can revolutionize efficiency and trust. For venture capital, this convergence will shift investment strategies. With AI-optimized analytics, VCs can better assess blockchain startups by analyzing decentralized data patterns and predicting long-term performance. Additionally, tokenized ecosystems powered by blockchain and AI will allow VCs to experiment with new funding models, such as staking, DAO-driven investments, and token-backed fundraising. In the next few years, these technologies will not only reshape industries but also democratize access to funding, fostering innovation at unprecedented levels. Let me know if you need any further elaboration or additional responses!

Thanks for sharing your knowledge and expertise. Is there anything else you'd like to add?

I’d like to emphasize that we’re just scratching the surface of what’s possible with the convergence of blockchain, Web3, and AI. These technologies have the potential to redefine how we approach trust, transparency, and scalability across industries, but the key to unlocking this potential lies in collaboration. Whether you’re a developer, entrepreneur, or investor, the future of decentralized technology will be shaped by those willing to innovate together.

At Crestal, we’re focused on building the tools and infrastructure to make this future accessible to everyone, but it’s a journey that requires constant learning, iteration, and feedback. If there’s one takeaway, it’s that the intersection of these technologies isn’t just about solving technical challenges—it’s about reimagining how we build systems that serve people and communities better.

Thank you for the opportunity to share my thoughts, and I look forward to seeing how this space evolves in the coming years!