Trevor Koverko has built his career focusing on areas where existing systems struggle to keep up with demand. His extensive work founding and leading companies across blockchain and artificial intelligence reflects a consistent focus on building underlying infrastructure rather than surface-level applications.
He became involved in blockchain in its infancy, including investing in projects such as Ethereum and Hashgraph. At a time when much of the attention was on trading activity, Koverko viewed blockchain as a way to improve how trust and coordination function across fragmented systems.
In 2017, he co-founded Polymath to address inefficiencies in capital markets. Traditional financial systems rely on intermediaries, manual processes, and slow settlement cycles. Polymath was built to enable the issuance of tokenized securities with compliance requirements built into their structures. The company raised tens of millions of dollars, generated eight-figure revenue, and contributed to early standards for security tokens.
That work later led to his co-founding of Polymesh, a blockchain designed specifically for regulated financial assets, with identity, governance, and compliance integrated into its architecture.
This same focus on addressing structural constraints is now guiding Trevor Koverko’s AI work, where data quality has become an increasingly important challenge.
A Constraint in Artificial Intelligence
While AI models and computing capabilities have advanced rapidly in recent years, the data used to train those systems remains a limiting factor.
AI systems depend heavily on human-labelled data to interpret language, context, and intent. Much of this work is produced through processes that can lack transparency and consistent quality control. For organizations building AI systems, particularly in regulated industries, this creates challenges around reliability and accountability.
Trevor Koverko has described this as a constraint that affects how far AI systems can progress. If the underlying data is inconsistent or difficult to verify, improvements in model design alone will not address the issue.
This ultimately led to his co-founding of Sapien.
A Different Model for Data Creation
Sapien is designed as a decentralized network that connects organizations with a global base of contributors who complete data labelling tasks, including annotation, classification, and evaluation work used in training AI systems.
The platform introduces incentives and accountability into the process. Contributors are rewarded for accurate work and stake tokens when completing tasks. If their work is judged to be low quality, they risk losing that stake. Tasks are reviewed by other contributors, and participants build reputation scores over time based on performance.
This structure creates a system where accuracy is directly tied to incentives. Contributors who perform well gain access to more opportunities and higher rewards, while those who perform poorly face penalties. The goal is to improve consistency and reliability in data production.
Trevor Koverko has emphasized that AI systems depend on the quality of their training data. Improving that data improves the system as a whole.
Bringing Transparency Into AI
A challenge in current AI development is limited visibility into how training data is created. Many organizations cannot fully trace the origin of their datasets or verify how labelling decisions were made, creating issues in sectors where accountability is required.
Sapien introduces traceability by recording contributor activity, review outcomes, and reputation within its network, providing organizations with more insight into how their data was generated and evaluated.
Advancing AI Data Infrastructure
Trevor Koverko’s expansion from blockchain to AI builds on his continued focus on creating modern, forward-thinking technology infrastructure.
As AI adoption continues to grow, the importance of high-quality training data is becoming more apparent. Larger models alone are insufficient without consistent, verifiable input.
Through Sapien, Trevor Koverko is working to improve how that data is generated and evaluated, with the aim of supporting the next phase of AI development.