Agentic Engineering
AI-assisted engineering workflows for monitoring, compliance support, developer tooling, automation, and technical operations.
- Builds agent-backed workflows that support expert judgement instead of replacing it.
- Explores compliance assistance, market monitoring, research automation, and operational tooling.
- Connects backend systems, structured data, and human-readable interfaces.
Motivation
Agentic engineering is the practical use of AI agents around real technical systems. The goal is not to make autonomous software that hides what it is doing. The goal is to build workflows where agents help engineers, operators, and domain experts see more context, move faster, and make better decisions.
This direction fits naturally around blockchain products, regulated platforms, and internal tooling. These environments produce many signals: code, logs, documents, requirements, market activity, compliance constraints, and operational state. Agents can help connect those signals when the underlying systems expose clear interfaces and stable data.
Engineering approach
The useful part is the system design around the agent. Domain logic should stay in ordinary software: APIs, services, databases, CLIs, reviewable rules, and explicit workflows. The agent becomes an interface and coordination layer over systems that can be tested, observed, and improved independently.
This approach keeps agentic tooling practical:
- agents operate through constrained tools and APIs;
- outputs remain inspectable by humans;
- important decisions keep a clear audit trail;
- backend systems own state and business logic;
- automation is introduced where it reduces repetitive work or improves monitoring quality.
Current directions
Current areas of interest include compliance support for regulated platforms, market monitoring, research and document analysis, developer workflow automation, personal infrastructure, and agent interfaces over backend systems.
The long-term direction is to make agents useful in engineering environments where correctness, security, and accountability matter.