Independent AI systems builder turning messy public data into trustworthy infrastructure for agents and people.
_{Kuala Lumpur, Malaysia 🇲🇾}

I build systems that make public data dependable: source discovery, freshness monitoring, schema and content validation, provenance and evidence trails, agent-ready contracts, MCP interfaces, signed attestations, and release/deployment verification. The work sits at the intersection of data engineering, AI agents, trust infrastructure, and production operations.

An open-source trust and interoperability layer for 418 official Malaysian public datasets. It provides a ten-status health taxonomy and continuously updated health snapshots; machine-readable dataset, licence, cadence, and provenance metadata; signed probe attestations and independently verifiable evidence receipts; and a read-only MCP server with 19 tools. llms.txt, agent.json, and mcp.json provide discovery surfaces, while fail-closed handling makes unavailable evidence explicit rather than guessing.

Live dashboard: data-pulse.my · MCP endpoint: mcp.data-pulse.my/mcp

The public documentation and sanitised agent-facing surface for a broader Malaysian data intelligence system. It documents generic data and evidence contracts, agent discovery files, a read-only MCP surface, and PDPA-clean sample data, with pharmaceutical compliance as the first operational vertical. The classified pipeline, scraping code, and private business systems remain separate.

• Trust layers rather than dashboards alone
• Agents as first-class consumers
• Full-system work, from contracts through operated releases
• Explicit uncertainty instead of implied certainty
• Prototypes turned into operated systems
• A Malaysian and Southeast Asian public-data perspective

• Start from the failure mode.
• Verify before claiming.
• Make uncertainty visible.
• Design contracts before integrations.
• Build once and verify the same artifact.
• Use AI agents for leverage while retaining architecture and acceptance responsibility.
• Turn repeated work into systems.
• Operate what ships.

• AI-agent and MCP architecture
• Data trust, freshness, provenance, and evidence
• Public-data engineering
• Pipeline and release reliability
• Machine-readable contracts and interoperability
• Practical AI adoption and workflow design
• Production-minded systems thinking

• Trustworthy AI and data systems for public-sector, regulated, and research use
• Malaysian and Southeast Asian public-data infrastructure
• Agent and MCP integrations
• Data engineering advisory
• AI adoption with governance and measurable outcomes

• Email: redzafahmy@live.com
• GitHub: r3dz4r
• DataPulse: live dashboard
• Kuala Lumpur, Malaysia

_{This profile README is maintained in the repository.}