
Urban Complaint Pattern Mining on NYC 311 Data (21M+ Records)
Data mining of 21.3M NYC 311 requests that found a six-complaint building 'syndrome' (mean lift 8.48) and validated it against 11M inspection records: 12-44x higher violation rates.
AI Engineer • Informatics Engineering Student
I build retrieval systems, fine-tune models, and ship the products around them. Currently building the RAG core of UniAI at Hasanuddin University.
Selected work in LLM systems, NLP, data science, and full-stack development.
View all projects→Where I have been applying this work.
Building the retrieval core of the university’s academic assistant: Qwen3 embeddings, BGE reranker, Qdrant and Redis caching, a 6-layer safety pipeline, and Qwen3-VL-8B served with vLLM on an NVIDIA L40S, load-tested toward 500 concurrent users.
Built an internal finance dashboard in a 4-person team (OPEX, cash advance, contract budget and vehicle monitoring), replaced manual spreadsheet reconciliation with reviewed imports, and deployed it to a Hetzner VPS after remediating every penetration-test finding.
Designed and built the cafe’s operating system solo: QR ordering, a Flutter staff tablet, an employee portal, and an owner dashboard on one Supabase backend.
Open to AI engineering roles, internships, research collaborations, and freelance projects.