PHAN VAN BANG

AI Engineer & Backend Developer

SUMMARY

AI Engineer with 2+ years of experience in designing, building, and deploying advanced AI solutions. Deep expertise in intelligent chatbots utilizing RAG, Agentic Systems, Tool Calling, and complex image processing workflows with ComfyUI. Practical experience in LLM fine-tuning, Prompt Engineering optimization, and AI infrastructure management from vector databases to cloud deployment. Always pursuing research and application of new technologies to optimize system performance and deliver real business value.

Short-term Goal: Take on an AI Engineer position at a forward-thinking enterprise, applying deep AI/ML expertise to develop breakthrough solutions, optimize data processing workflows, and contribute directly to the success of high-value products.

Long-term Goal: Develop into a leading expert in the AI field, spearheading pioneering technology initiatives and building strong AI teams, contributing to shaping the future of artificial intelligence in practical applications.

TECHNICAL SKILLS

Programming Languages:
Python, JavaScript & TypeScript, C#, Java, C++, C
Frameworks & Libraries:
LangChain, LlamaIndex, FastAPI, Flask, Streamlit, Spring Boot, Laravel, ExpressJS, React, React Native, Unity
Databases:
Vector DB (Qdrant, Faiss, Chroma), SQL (MySQL, SQLite), NoSQL (MongoDB, Firebase), Graph DB (Neo4j)
AI/ML & Image Processing:
LLM, Prompt Engineering, LLM Fine-tuning, RAG, Agentic Systems (Tool Calling, CrewAI, Model Context Protocol - MCP), Tesseract, Diffusion Model, ComfyUI Workflows, HuggingFace, vLLM, Ollama, LlamaCPP
Architecture & Systems:
Microservices, RESTful API Design, Docker, VPS Management (Ubuntu, Linux), Google Colab, Jupyter
DevOps & CI/CD:
AWS (S3, EC2), Cloudflare, Vercel, GitHub Actions
Tools & Processes:
Git, N8N Automation, Data Crawling/Scraping, Unit Testing

WORK EXPERIENCE

AI Engineer | Backend Developer

06/2023 - Present
MekongAI
AI Solutions Development and Deployment:
  • Implemented Retrieval-Augmented Generation (RAG) architecture for chatbot systems, improving information retrieval accuracy by 20-30% and reducing average response time to under 2 seconds, enhancing user experience.
  • Participated in building AI systems based on Agentic Systems (Tool Calling, CrewAI, Model Context Protocol - MCP)
  • Fine-tuned and optimized Prompt Engineering, increasing LLM problem-solving success rate by 30% and ensuring consistency and relevance of AI-generated content.
  • Participated in developing and deploying specialized AI chatbots (Law for Prosecutor's Office, Social Media Facebook, Vehicle, etc.) using LangChain, LLM, and architectures like RAG and Agent to solve client problems, receiving positive feedback on quality and usefulness.
AI Model Training and Optimization:
  • Performed Fine-tuning of LLMs on datasets, enabling models to acquire new knowledge and respond more appropriately to context.
  • Trained LoRA models (Generative AI) using AI-Toolkit, Flux on image datasets, adjusting appropriate parameters to enable high-fidelity product image generation and refinement.
Backend Development and Database Management:
  • Designed and developed high-performance RESTful API endpoints using FastAPI, maintaining average response time under 200ms to efficiently serve AI backend services and frontend applications.
  • Deployed and optimized Vector Database (Qdrant) for efficient storage and querying of vector embeddings, and Graph Database (Neo4j) for knowledge graph construction.
  • Managed and optimized SQL (MySQL) and NoSQL (MongoDB) databases, ensuring data integrity and stable query performance for critical applications.
AI Image Processing Technology (Diffusion Model, ComfyUI):
  • Built and managed over 20 complex image processing workflows using ComfyUI, including applications for Image Generation, FaceSwap, OutfitSwap, Background Replacement, Image Restoration, DeepFake, Virtual Try-On (VTO), etc.
  • Developed WebSocket API to connect workflows with ComfyUI, enabling processing API creation and result delivery to Frontend.
Infrastructure Management and Automation:
  • Administered VPS (Ubuntu, Linux) for deploying tasks such as Local LLM, Training, Testing, etc. for AI and backend applications.
  • Used N8N Automation to automate over 5 repetitive workflows (e.g., RAG, integration with applications like Google Calendar, Drive, Zalo, Discord, Telegram, data collection and processing, etc.) reducing manual task time by 40%.
  • Deployed and managed cloud hosting with domain management (Hostinger), configured subdomains via Cloudflare to optimize access performance and enhance security for project web applications.
  • Managed AWS S3 for efficient data storage and retrieval, including documents, images, and static resources for AI systems, ensuring high scalability and reliability.

EDUCATION

Engineer of Information Technology

2020 - 2025
Vietnam-Korea University of Information and Communication Technology (VKU) Da Nang City, Vietnam

SOFT SKILLS

Teamwork & Effective Collaboration: Experienced in working and collaborating closely in team environments, proactively sharing knowledge and supporting colleagues to achieve common project goals.
Coordination and Team Development Experience: Experienced in supporting small development team coordination, including participating in task assignment, progress tracking, and team member guidance to ensure smooth collaboration and efficient achievement of common project goals.
Analytical Thinking & Problem Solving: Approach technical challenges systematically, analyze root causes deeply, and propose creative, optimal solutions that deliver practical results.
Communication & Proactive Collaboration: Communicate technical information clearly and coherently with team members (including non-technical members) and stakeholders, promoting mutual understanding and effective collaboration.
Learning Spirit & Continuous Development: Proactively update knowledge and skills in new technologies, especially in AI and software development fields, ready to experiment and apply advanced solutions to enhance personal capabilities and contribute to team development.