# Van‑Loc Nguyen (Nguyễn Văn Lộc) — Felix

> A computer‑vision and deep‑learning engineer working at the seam between research and production.


## About

My name is Nguyễn Văn Lộc (you can call me Felix) - a professional and enthusiastic programmer with a particular love for Artificial Intelligence, especially Computer Vision and Deep Learning. I am a quick learner with a self-learning attitude, always eager to explore new technologies and passionate about solving hard problems. Python is my core language and closest tool; I love almost everything that can be built with it.

Over the last several years I've moved between a research lab at the University of Science, VNUHCM, a backend engineer role at Saritasa, an AI engineer role at Autonomous, a student researcher position at the National Institute of Informatics in Tokyo (Echizen Laboratory, under Prof. Isao Echizen), and now a Software Engineer position at Eurofins. The throughline is the same: take a model out of the notebook and into something a person can use without thinking about it.

I read widely, ship steadily, and believe that the most interesting problems sit between disciplines — vision and language, research and product, Vietnamese and Japanese, theory and the messy thing it's supposed to model.

## At a glance

- **Current role:** Software Engineer, *Eurofins*
- **Now researching:** MSc · *AI*, HCMUS
- **Open to:** — Collaboration · Talks
- **Recently:** Returned from *NII, Tokyo*
- **Based in:** Ho Chi Minh City

## Now

- **Shipping:** [ArcanaAI](https://arcanaai.nguyenvanloc.com)— redesigned UI with smoother tarot flow, clearer card storytelling, and a more immersive mystical aesthetic. (Personal project · Live)
- **Researching:** Zero‑shot object removal and multi‑object reasoning for video, continuing work from PANDORA and Cluster‑Erase. (Master's thesis · HCMUS)
- **Reading:** Diffusion world models, evaluation of vision‑language systems, and the long quiet literature on agents that don't hallucinate. (Ongoing)

## Selected work

### 01 · Doodle Duel (Jankenpon)
A real‑time competitive web party game where a vision‑LLM acts as referee on a shared canvas.
- Tags: Live, React, Gemini API, Socket.IO
- Link: https://github.com/heyGio/jankenpon/tree/loc/hehe

### 02 · ArcanaAI
AI‑powered tarot reading service — OpenAI function‑calling, VertexAI, persistent state with PostgreSQL & Redis, and a deliberately mystical UI.
- Tags: Live, Next.js, OpenAI, VertexAI, Redis, Supabase
- Link: https://github.com/vanloc1808/arcana-ai

### 03 · Fake News Analysis & Detection
End‑to‑end neural networks for cheapfake detection, paired with online reputation checking and real‑time authenticity guidance. Undergraduate thesis project.
- Tags: PyTorch, Streamlit, Docker
- Link: https://github.com/nbtin/fakenews-detection-demo

### 04 · Connect
Feature‑rich real‑time chat app — group chat, file sharing, and a full CI/CD pipeline — built for the Software Architecture course.
- Tags: ReactJS, Django REST Framework, PostgreSQL
- Link: https://github.com/software-architecture-20tn/chat-app-frontend

### 05 · MiniShopping
E‑commerce Android app with catalogue, cart, order processing, and payment integration on a Django REST Framework backend.
- Tags: Android, Django REST Framework, NoSQL
- Link: https://github.com/mekanican/MiniShopping

## Selected publications

### Cluster Erase: Zero‑Shot Mass‑Similar and Multi‑Object Removal in a Single Pass
IEEE International Conference on Multimedia & Expo Workshops — ICMEW 2026, Bangkok · Accepted · May 2, 2026

A demonstration of a single‑pass removal approach for visually similar object clusters in natural imagery, extending zero‑shot inpainting toward the messy, repetitive scenes that one‑object methods choke on.

### PANDORA: Pixel‑wise Attention Dissolution and Latent Guidance for Zero‑Shot Object Removal
IEEE International Conference on Multimedia & Expo — ICME 2026, Bangkok · Accepted · Mar 17, 2026

A zero‑shot object removal method that leverages pixel‑wise attention dissolution and latent guidance within a diffusion framework — achieving clean inpainting without task‑specific fine‑tuning on the target scene.

### Visionary: Optimized Temporal Video Retrieval via Large Language Model‑Enhanced Query Processing
The 14th International Symposium on Information and Communication Technology (SOICT 2025) · Accepted · December 13, 2025

Addressing the Ho Chi Minh City AI Challenge 2025, Visionary introduces an adaptive keyframe extraction algorithm, a Qwen3‑VL pre‑processing pipeline for metadata generation with integrated OCR, a flexible architecture supporting multiple embedding models, and Reciprocal Rank Fusion for synthesizing retrieval results across complex, large‑scale video tasks.

### ENAug: ENT Endoscopy Images Classification Using Ensemble and Augmentation Methods
The 14th International Symposium on Information and Communication Technology (SOICT 2025) · Accepted · December 13, 2025

A practical pipeline for classifying ENT endoscopy imagery that pairs an ensemble of vision backbones with a curated augmentation schedule, designed to remain robust to the imaging variance typical of clinical capture devices.

## Recent journal

- **May 2, 2026** · Paper — Demo paper *Cluster Erase: Zero‑Shot Mass‑Similar and Multi‑Object Removal in Single Pass* accepted to the IEEE ICME 2026 Workshop in Bangkok, Thailand. (ICMEW · TH)
- **Apr 1, 2026** · Role — Started as *Software Engineer* at Eurofins. (Eurofins · VN)
- **Mar 28, 2026** · Move — Left the role of *Student Researcher* at the National Institute of Informatics (NII), Tokyo, Japan. (NII · JP)
- **Mar 17, 2026** · Paper — Paper *PANDORA: Pixel‑wise Attention Dissolution and Latent Guidance for Zero‑Shot Object Removal* accepted to the 2026 IEEE International Conference on Multimedia and Expo (ICME) in Bangkok, Thailand. (ICME · TH)
- **Feb 21, 2026** · Event — Participated in the *Gemini 3 Hackathon Tokyo* (Gemini 3 ハッカソン 東京), hosted by Cerebral Valley and Google DeepMind. (Tokyo · JP)

## Elsewhere

- GitHub — vanloc1808: https://github.com/vanloc1808
- Google Scholar: https://scholar.google.com/citations?user=vanloc1808
- LinkedIn: https://linkedin.com/in/vanloc1808
- Stack Overflow: https://stackoverflow.com/users/vanloc1808
- Facebook: https://facebook.com/vanloc1808