AI Engineer turning multimodal research into production. Production AI systems, multimodal retrieval, OCR, RAG, MLOps, and event-driven backends in Go and Python.

I’m an AI Engineer based in Ho Chi Minh City, working at the intersection of multimodal retrieval, OCR, and production ML systems. I turn research-grade ideas into shipped services — Kafka, Cassandra, FastAPI, and Go around the model, not just the model itself.

What I have shipped recently:

• Top-4 @ EVENTA Challenge, ACM Multimedia 2025 — first author of ZSE-Cap (mAP 0.966, R@1 0.955) on the 400K+ OpenEvents V1 dataset, zero fine-tuning.
• Best Presentation Award @ SOICT 2025 — co-author of KPTER, an $O(K \cdot N)$ K-pointer algorithm for temporal multi-event video retrieval that replaces $O(N^K)$ nested loops.
• +30 points H-mean on Sino-Nom OCR (0.731 → 0.952) — fine-tuned PP-OCRv5 with knowledge-distilled PP-HGNetV2_B4 on a 3.4K+ historical-manuscript corpus I curated with a 16-Goroutine scraper.
• Backend Engineer (AI Platform) @ Acamar AI Vietnam — building Go and Python microservices, event-driven Kafka pipelines, and Dockerized inference workers across GCP, AWS, and Azure.

I’m pursuing a B.Sc. in Artificial Intelligence at the University of Science, VNU-HCM (GPA 8.79/10.0). Outside of papers and PRs, I prototype small Go tools, write up what breaks, and read more than I push to GitHub.