Explain public Ethereum activity without taking control of user assets.

Blockchain analysis becomes a trust problem as soon as AI enters the workflow.

The platform retrieves and organizes public Ethereum activity, gives users a plain-language analytical interface, and uses AI to help interpret addresses and transactions. An analytical product must not quietly acquire authority over assets or blur the line between information and execution.

That makes data boundaries, authentication, long-running work, auditability, database evolution, and repeatable delivery part of the product design rather than supporting infrastructure.

Build the product foundation around explicit boundaries.

Application and data foundation

I helped shape the platform around Python and Flask, with SQLAlchemy and Alembic providing a controlled path for data access and schema evolution across PostgreSQL environments.

Ethereum and AI orchestration

The system integrates Etherscan for public Ethereum data and OpenRouter for AI-assisted interpretation. Service boundaries and background jobs keep external calls and longer-running analysis separate from the interactive flow.

Authentication, audit, and delivery

Google OAuth supports authenticated access. The product remains read-only and non-custodial, while append-only audit controls preserve a trace of relevant activity. CI and deployment workflows make database and application changes repeatable.

• Python
• Flask
• SQLAlchemy
• Alembic
• PostgreSQL
• OpenRouter
• Etherscan
• Google OAuth
• Background jobs
• Audit controls
• CI
• Deployment

A foundation that can evolve without weakening its trust model.

The result is a backend foundation that connects public Ethereum data and AI-assisted analysis while keeping the core boundary explicit: the product observes and explains, but does not take custody of funds or keys. Database migrations, background processing, audit controls, and delivery automation make that boundary maintainable as the platform evolves.