***********.****@*****.***
https://github.com/ducmanh1312
Jun 2024 - Mar 2025
Intelligent Connected Network Lab, PTIT
AI Research & Python programming
TOEIC
Language: Python
Analyst data and get business insight
May 2024 - Aug 2024
Global Technology Solutions, JSC
Data & Business analyst
(course material)
Deep learning specialization
Mar 2025 - July 2025
ProxGlobal
AI Engineer
Florant - AI Plant classification
ProxGlobal
AI beauty
ProxGlobal
Research on AI model and optimize technique
Collaborate with team to release paper about SOTA AI technique
Program language
Python: TensorFlow, Pytorch, Scikit-
learn, Langchain, Langgraph
C++, HTML/css, Typescript
Database: SQL, MongoDB, Qdrant, ...
Deploy: BentoML, FastAPI, MLflow
Docker
Version control: Github, Gitlab
Software development: Agile-scrum
workflow
POSTS AND TELECOMMUNICATIONS
INSTITUTE OF TECHNOLOGY (PTIT)
Major: Electronics and
Telecommunications Engineering
2021-2024
NGUYỄN ĐỨC MẠNH
AI intern - fresher
Apply the knowledge I have learned, continue
learning during the internship to develop myself
to become a professional AI engineer
Join hands to creating AI products that help in
the development of your company
Education
Skills
Certification
Experience
Project
GPA: 3.2
Key responsible in 2 AI project for mobile app
Collaborate with other departments to design AI workflow compatible with business objective
Train AI model & Integrate to current workflow
Key responsible:
Research, test, propose compatible AI model and workflow Train model using one-shot learning techniques
Optimize vector database for fast retrieval and efficient payload storage
Apply AI technique to enhance accuracy ( ~90%)
Deploy to production & Optimize performance for multi user request
Key responsible:
Research, test, propose compatible AI model
Deploy multi-modal for each function:
Inpainting; remove object, watermark, people
Diffusion models: replace object, outpainting
Modify text, replace text
Test, optimize generative parameter proper with production Huggingface LLM course
(link)
A face eKYC system for secure and efficient identity verification. Face eKYC
Key feature:
Face Detection: Integrated a YOLOv8-based model with for real-time face detection and estimate head pose
Implement Facenet and ArcFace for face recognition Training and deploy Liveness detection to identify real live person
User Interface & Backend:
GUI: Streamlit integrate with FastAPI and MongoDB database Real-time data retrieval from camera with RTSP protocol Key Features:
Fine-tune model to remove the background of the invoice. Utilize CRAFT for text detection and combine it with EfficientNet model to rotate the image.
Fine-tune VietOCR to invoice text recognition.
Apply PhoBERT & SVM to classify extracted text into structured fields Design UI for real-time interaction and visualization Invoice OCR
An OCR systerm to extract, classify information from Vietnamese invoices
(link)
Chatbot:
Implement LLM agent that dynamically integrates multiple tools: Text- to-SQL, Google Search, document RAG
Integrate session-based and message history management GUI: Streamlit, Sqlite database, packaged by Docker Advance document RAG:
Document retrieval: Hybrid Search, Re-ranking, and Self-Query Retriever Manage multimodal to extract image, text, table from documents and generate response
Apply quantization techniques to optimize transformer model Advance RAG chat bot (link)
My chatbot assistant with document RAG and Multi-tool integration Personal project
Design IOT system and network
connection protocols
IOT system design certificate
Project management
certificate