Vu Nhat Minh

Applying for: Full-stack Engineer - Junior AI Engineer

Education

FPT UniversityHoa Lac, Hanoi

Sofware Engineering09/2022 - Now

Full-stack Developer InternFPT Software Academy (FSA)May 2025 - Aug 2025

Built sachify.id.vn, a full-stack manga/story reading platform, using ReactJS for the frontend and ExpressJS for the backend API
Designed and implemented the database schema and integrated it with backend services to support content management and user reading features
Styled responsive UI components with Tailwind CSS and Bootstrap, ensuring consistent experience across devices

Programming Languages: C++, JavaScript / TypeScript, HTML / CSS, Java, Python, SQL, Dart

Frontend: ReactJS, NextJS, Flutter

Backend: NodeJS, ExpressJS, RESTful API, Spring Boot, JSP Servlet

Database: PostgreSQL, MySQL, MongoDB, SQL Server

Tools and Frameworks: Git, Linux, Postman, Coolify, Google Cloud Platform, VS Code, Cursor, Antigravity, Claude Code, Codex

English: Proficient in professional communication and writing

Japanese: Basic (elementary vocabulary and grammar, JLPT N5 in progress)

Spring Boot, Next.js, PostgreSQL | edua.memore.vn | OpenAI, Deepseek

Contributed to an AI-powered platform for high school teachers, supporting lesson plan creation, slide editing, virtual labs for Physics, Chemistry and Math subjects
Architected interactive physics simulation engines for the Virtual Lab module aligned with the national curriculum, optimizing client-side rendering performance.
Authored system architecture specs (SRS/SDS) and designed RESTful APIs with role-based access control and Google OAuth 2.0.
Containerized and deployed production services to VPS via Coolify, managing reverse proxy, custom domain DNS, and CI/CD workflows.
Integrate Deepseek v4 flash model for text content generate and OpenAI GPT-Image 2.0 for image content generate.

Python, PostgreSQL | Claude, Deepseek, OpenAI, Gemini

Designing and developing Rei, a personal AI agent with persistent memory, daily wellness/task tracking, and conversational reminders/note-taking, using a model-agnostic architecture that decouples persona and memory to support future LLM upgrades without data loss.
Developing extensions for tool-calling based device automation, Discord bot integration, and voice interaction (speech-to-text/text-to-speech).
Intergrate Claude Haiku 4.5 and Deepseek v4 pro for thinking, reasoning, prompt writing, tool calling, and generating responses; integrate OpenAI GPT-4o and DeepSeek v4 Flash Vision for image reading and scanning tasks; and integrate Gemini 3.1 Flash TTS for voice capabilities.

React, Firebase, Google Cloud Functions | syncho.site

Independently designed and built Syncho, a real-time collaborative workspace enabling users to securely access their data from any device using a simple ID and password, without requiring account registration
Integrated Tiptap rich text editor for note-taking and image storage, alongside Kanban boards and a whiteboard tool, creating a unified multi-purpose workspace
Built serverless functions on Google Cloud to handle backend logic, paired with Firebase Firestore and Authentication for real-time data sync and secure, ID-based access control across devices