Portfolio
Minji Kim
Enterprise AI & Automation Consultant @ VIPC · Brumley Graduate Fellow @ Strauss Center
About
About Me
Hello, I am Minji! I work at the intersection of people, technology, and information, dedicated to creating systems and designs that work for people. My focus is on building solutions that promote equitable access to resources, and I work on procuring and deploying technology that is accountable, inclusive, and built for the communities it serves to expand opportunities.
Currently, I am an Enterprise AI & Automation Consultant at VIPC, where I lead AI tool intake and approval and deploy AI agents that automate document generation, validation, and compliance review across divisions. I am also a Brumley Graduate Fellow at the Strauss Center for International Security and Law, researching governance for cross-border federated AI in financial crime detection, and an M.S. in Business Analytics candidate at UT Austin's McCombs School of Business.
BS Human-Centered Data Science (GPA: 3.9) @ UT Austin
Minor: Risk Management, Programming & Computation
Graduate Brumley Fellow @ Strauss Center
Research & Work
Projects & Publications
- Introduced a novel dataset and vision-language model baseline for rewriting visual instructions while preserving user privacy, advancing responsible deployment of multimodal AI.
- Conducted as part of the AI/ML/NLP Labs research experience at UT Austin (Aug 2024 – Sep 2025).
- Supported by the UT Austin Undergraduate Research Fellowship ($7,500 in funding).
- Leading comparative analysis of AML information sharing, data protection, and AI oversight requirements across 4+ jurisdictions.
- Designing a federated learning proof-of-concept that benchmarks independent, centralized, and federated AML models to surface governance trade-offs.
- Proposing an interoperable governance framework covering accountability, model auditability, privacy, and regulator access, aligned with financial sector cybersecurity standards in partnership with the Cyber Risk Institute.
- Built a vector-powered research intelligence platform on AWS processing 200K+ publication records.
- Maps researcher pipelines and identifies commercialization opportunities across Virginia's university and higher education labs.
- Bridges the gap between academic innovation and market pathways for Virginia's startup ecosystem.
- Analyzed 4,000+ property records with Pandas and Matplotlib to identify property deterioration patterns across Pittsburgh.
- Filled gaps in the original dataset by consolidating 100K+ external records, then validated it for risk modeling.
- Joined property and geospatial data into interactive Power BI maps of deterioration hotspots, presenting mitigation recommendations to the city's Chief Data Officer.
- Designed a multimodal emotion recognition system combining speech audio and ASR-derived text.
- Trained a late-fusion model on 7,400+ audio files, boosting classification accuracy by 30% over unimodal baselines.
- Built a prototype processing live voice inputs to evaluate real-time model performance.
- Built a phishing website detector trained on 10,000+ URLs, reaching 98% accuracy after comparing 6+ classification models.
- Narrowed 50 features to 13 to make model decisions explainable and reduce false positives that would block legitimate sites.
- Designed a real-time URL checker that lets non-technical users flag suspicious links before clicking, refined through usability testing.
Curriculum Vitae
Experience & Education
Education
Minor: Risk Management, Programming & Computation
Experience
AI tool intake and approval · AI agents across 4+ divisions · JLARC validation agents (20% of workflow automated, 30% faster review)
Governance of cross-border federated AI for anti-money laundering
Responsible AI and data governance policy · AI governance controls · AWS research intelligence platform · AI chatbot (500+ monthly users)
AI analytics platform tracking 2,000+ companies · legacy system consolidation saving 7,000+ hours annually
NIST-based assessment, security governance policies, and City Council briefing for Devine, TX
AI adoption across 3+ divisions · demand forecasting (10% lower inventory costs)
Leadership
Honors & Involvement
Skills
Coursework
Courses
Certifications
Foodie Diary
Cafe Reviews
Suggest a Café
Know a spot I should try? Fill in the form and I'll check it out!
Contact
Let's Connect
Whether you're interested in collaborating on research, talking data and AI, or just want to swap café recommendations — I'd love to hear from you!