Rodrigo Coelho

Rodrigo Coelho is an entrepreneur and blockchain-infrastructure executive who became The Graph’s first employee and was later appointed chief executive of Edge & Node. His work connects the data applications need to make decisions with the financial controls required when AI agents can act on those decisions and spend money.

From building businesses to building infrastructure

Coelho began with an application-development business in the early web era, then founded and exited two technology startups before joining The Graph. His entrepreneurial progression brought together an industrial-engineering background and experience running technology companies. He had met Graph co-founder Yaniv Tal in a San Francisco coworking space before joining the team, moving from developing individual businesses to helping build infrastructure other developers could use.

The Graph addresses a practical problem: blockchain transactions need indexing before applications can efficiently retrieve useful information from them. Coelho helped build the organization around that technical work, managing operations, developing its community, and onboarding employees. He personally helped colleagues join the company before later becoming their chief executive. As CEO, he reduced costs and refocused the organization around shipping new products.

His earlier role as Chief Spirit Officer emphasized creativity, purpose, and connection within the team. His approach to leadership paired attention to organizational culture with removing obstacles that prevent people from shipping. His contribution extended beyond the underlying protocol to the people and working conditions needed to sustain its development.

Connecting usable data with authorized action

Edge & Node’s newer products extend its blockchain experience into institutional data and agent commerce. Amp turns blockchain activity into structured blockchain datasets accessible through SQL, with cryptographic provenance tracing information to its source. It handles chain reorganizations and historical backfills, allowing applications to work with changing blockchain records without managing every underlying indexing task themselves.

Ampersend addresses the spending side. It evaluates agent payments against budgets, merchant restrictions, and approval thresholds before funds move, while retaining transaction records. Reliable data helps an agent decide what to do; payment policies determine whether it is authorized to do it.

Coelho connects this direction to Edge & Node’s earlier query micropayments: purchasing small units of blockchain information becomes relevant when agents need to buy data or tool access as part of a task. The work also encompasses payment batching to reduce blockchain transaction fees when individual payments are extremely small. This gives agent commerce a concrete starting point in an existing infrastructure problem—how software pays for the information it consumes.

What accountability requires

• Machine-speed payment controls. Coelho argues that agents operating continuously need financial policies enforced within the transaction workflow. Procedures that depend on a person reviewing each action cannot simply be carried over to software transacting around the clock. His work on agent spending, presented jointly with Pranav Maheshwari, places budgets, authorization, and compliance alongside the ability to move money. Human legal and policy responsibility remains in place even when execution is automated.

• Payment access and policy enforcement are separate capabilities. Ampersend connects agents to paid information tools, allowing a small payment to unlock an endpoint needed for a task. But a functioning checkout alone cannot guarantee that an agent respects its instructions. A gift-purchase example in the joint presentation pairs a request to spend less than $10 with a reported $9 total, an $11 charge, and a $10 gift. The recording does not resolve these amounts or establish that the final charge obeyed the requested limit. This uncertainty makes the engineering requirement tangible: spending controls must govern the complete amount charged, rather than merely the price of the selected item.

• Counterparty screening before funds move. A wallet address alone does not tell an enterprise who it is paying or whether the recipient is sanctioned. Ampersend’s TRM Labs integration checks counterparties before payment execution. In the joint presentation’s screening simulation, an ordinary wallet and a simulated flagged wallet can both pay for a scraping service while screening is disabled; enabling screening rejects the flagged wallet’s payments. This illustrates a specific preventative mechanism within Coelho’s broader argument for financial governance, while leaving other transaction and compliance risks to their own controls.

• Enterprise AI accountability. In his writing on enterprise AI security, Coelho argues that automated decisions must be explainable to auditors, boards, customers, and regulators. Fragmented logs and inconsistent transaction records make that difficult. He favors structured data, checks before execution, and inspectable records that let people understand what happened and why. Accountability depends on the systems surrounding an agent as well as the model itself.

Coelho’s approach to decentralized infrastructure emphasizes useful developer tools, competitive performance, verifiable information, and data portability. Across The Graph, Amp, and Ampersend, his focus has developed from making blockchain information retrievable to making it usable by institutions and governing the software that acts on it.