Hero

Enterprise AI Governance Copilot is a full stack governance platform built to help organizations scale AI responsibly through clear policy, risk, and compliance controls.

Problem

As AI adoption expands, teams often lose visibility across policies, approvals, and compliance status, creating governance risk and weak auditability.

Solution

I built a governance dashboard and backend platform that centralizes policy lifecycle workflows, risk tracking, role-based accountability, and audit trails in one operating layer.

Core Features

• Policy management with versioning and approval workflows
• Compliance dashboard with risk scoring and status visibility
• Role-based access control for governance responsibilities
• Audit logging and activity traceability
• Monorepo structure for coordinated frontend and backend development
• NestJS REST API backed by Prisma ORM

Governance Workflow

The governance workflow coordinates repository ingestion, multi-domain scanning, risk scoring, and recommendation generation so teams can move from raw findings to executive action in one operating loop.

Technical Architecture

Implemented in a Turborepo monorepo with a Next.js frontend and NestJS backend. PostgreSQL and Prisma manage policy, compliance, and audit activity with strongly typed data models.

Monorepo Architecture

The system architecture separates dashboard, governance API, scanning pipeline, scoring services, and policy configuration while preserving type-safe data access and shared operational visibility.

Key Technical Decisions

• Used a monorepo strategy to share types and speed coordinated development
• Modeled governance ownership with structured RBAC boundaries
• Designed API-first service layers to support future enterprise integrations

Engineering Challenges

• Modeling policy and compliance workflows without introducing unnecessary complexity
• Designing risk views that are meaningful to both technical and non-technical stakeholders
• Maintaining clean domain boundaries between dashboard and backend services

Results and Impact

The platform creates centralized governance visibility and gives enterprise teams a practical foundation for compliant, repeatable AI operations at scale.

Key Skills Demonstrated

Enterprise architecture, governance workflow design, API engineering, monorepo implementation, and policy-aware product design.

Lessons Learned

Governance systems succeed when information architecture, role clarity, and traceability are as strong as the technical implementation.

Hero

Guesthouse Accounting System is an offline-first accounting tool created for real daily operations and designed specifically for a 74-year-old non-technical operator.

Problem

Manual spreadsheets were slow and error-prone for daily bookkeeping. The existing process was too complex for non-technical use and unreliable in low-connectivity conditions.

Solution

I built a senior-friendly interface that streamlines transaction entry, persists records locally, and provides clear exports for accounting review without requiring internet access.

Core Features

• Simplified transaction capture optimized for non-technical daily use
• Accessibility-focused UI with large readable inputs and clear flow
• Offline-first architecture with local JSON persistence
• Excel export for monthly reconciliation and external accounting handoff
• Structured income and expense views for faster bookkeeping

Technical Architecture

Implemented with HTML, CSS, and vanilla JavaScript, using local persistence with JSON-style data handling and export utilities for spreadsheet-compatible reporting.

Local Data Persistence

The system keeps guesthouse records locally so daily operations continue without an internet connection while still preserving a simple, structured storage model for later export.

Senior-Friendly UX / Accessibility

The interface was kept intentionally simple with large controls, clear labels, and a low-friction workflow so the operator can complete tasks confidently without navigating complex accounting software.

Transaction and Reporting Workflow

Transaction capture, invoice generation, and reporting were designed as a straightforward operating loop so the user can record activity first and produce accounting outputs when needed.

Key Technical Decisions

• Prioritized interaction simplicity over feature density for consistent adoption
• Designed around accessibility and low-friction keyboard/mouse usage
• Chose offline local storage patterns to remove internet dependency from core tasks

Engineering Challenges

• Balancing accounting structure with an interface suitable for a senior non-technical user
• Keeping local data resilient while maintaining straightforward export workflows
• Reducing cognitive load without compromising bookkeeping accuracy

Results and Impact

The system replaced high-friction spreadsheet work with a reliable offline workflow, improving speed, consistency, and confidence in day-to-day financial capture.

Key Skills Demonstrated

Human-centered systems design, accessibility-first UX, offline architecture, local data handling, and practical business process digitization.

Lessons Learned

Internal software delivers real value when it adapts to the user, not the other way around. Simplicity and accessibility drive lasting operational adoption.