I'm a Strategic Change Manager & Master's Student in Industrial Analytics at Uppsala University. I bridge the gap between business strategy and technical implementation, with a focus on AI Adoption and Rapid Prototyping.

Here are the demos of the MVPs I managed and designed during recent hackathons.

Role: Product Strategy & Compliance Logic | Time: 30-hour Hackathon Tech Stack: Ollama (Local LLM), Python, Regex, Google Gemini, Anthropic Claude, Lovable Focus: AI Safety, Data Sovereignty, GDPR/EU AI Act Compliance

Description: A bi-directional privacy firewall designed to solve the "Enterprise AI Dilemma" (using Cloud AI vs. keeping data safe).

• The Problem: Addressed OWASP LLM06 (Sensitive Information Disclosure) where employees accidentally leak PII to public LLMs.
• The Solution: Implemented "Reversible Anonymization". The system detects and replaces sensitive data (Names, IDs) with tokens locally before sending the prompt to the cloud, then reconstructs the answer locally.
• Strategic Impact: Enables "Compliance as a Service," allowing regulated industries (Finance, Healthcare) to adopt GenAI while maintaining absolute data sovereignty.
• Project Demo: Watch our pitch and technical demo here: https://www.youtube.com/watch?v=pBuz-j09LYA

Role: Team Lead & Product Design | Time: 90-minute Sprint Tech: Lovable.dev (GenAI), Web App

Description: A social hub for international students featuring an "Interest Match" algorithm and a gamified "Daily Event Blind Box".

Check out the demo of the project: https://github.com/juneisxiaoliu/juneisxiaoliu/blob/main/After%20Class%20Demo.mp4

Role: Team Coordinator | Time: 24-hour Hackathon

Description: An AI-powered tool streamlining information retrieval for university events.

Check out the demo of the project: https://github.com/juneisxiaoliu/juneisxiaoliu/blob/main/team%202%20loveable%20event.mp4

Role: Product Strategy & UI Architecture | Time: 48-hour Hackathon

Description: An enterprise compliance platform designed to resolve the AI governance gap in cross-border corporate expansion. The Problem: Addressed the "Compliance Double Bind" where business teams leak strategic data via Shadow AI, while legal teams reject LLMs due to hallucinations and zero liability.

The Solution: Implemented "Intent-Preserving Guardrails" to locally redact financial magnitudes (e.g., budgets) before API calls, paired with a GT VERIFIED mechanism that grounds AI output in human-audited legal databases.

Project Demo: Watch our pitch and technical demo here (https://drive.google.com/file/d/166cdxC4YBskv1dLSBP0VBcL580-fEkHU/view?usp=sharing)