Ai Engineer Ho Chi

WORK EXPERIENCE

SUMMARY

AI Engineer Encycom ** District,Ho Chi Minh City January 2025 - Present Engineered and deployed an automated data crawling system using N8N and Crawl4AI to extract and process high-volume order data from e-commerce platforms (Amazon, TikTok), creating a clean and structured data pipeline for internal analytics. resulting in a 95% reduction in manual processing time and enhanced analytics efficiency

A dynamic and innovative AI Engineer with a proven track record of developing and deploying sophisticated AI systems from concept to production. Specializing in Vietnamese Natural Language Processing (NLP), I possess deep expertise in building complex applications, including voice-activated assistants and intelligent search engines using Retrieval-Augmented Generation (RAG). My proficiency in modern MLOps practices—including Kubernetes, Docker, Triton Inference Server, and the ELK Stack— ensures that the solutions I build are not only intelligent but also robust, scalable, and maintainable. I am passionate about tackling complex challenges and am driven to apply my skills in LLMs, data engineering, and system optimization to create high-impact AI solutions that deliver tangible business value. AI ENGINEER

Address:

Phone:

Email:

Website:

36 Street, Linh Dong Ward,Thu Duc District, Ho Chi Minh City 034*******

*********@*****.***

https://github.com/Haole3435

Project: Automated E-commerce Data Pipeline

Project: AI-Powered Design Verification Tool

Developed a computer vision tool to automatically validate product design images against complex customer requirements, significantly improving quality assurance efficiency and ensuring design compliance.

Managed the containerization of AI services with Docker and contributed to deployment strategies, ensuring scalability and maintainability in a production environment. AI Engineer Illuminous AI Binh Thanh District,Ho Chi Minh City May 2024 - November 2024 Project: Virtual Try-On using Generative AI (StyleGen) Spearheaded data acquisition by crawling over 20,000 images from 50 leading fashion brands to build a comprehensive, high-quality dataset.

Enhanced the core inpainting algorithm by refining the masking method, which resulted in a 27% measurable improvement in final image quality.

Drove model optimization for production by converting the Stable Diffusion UNet to the TensorRT format, achieving a 9x increase in inference speed. Packaged the final, optimized model as a production-ready service, exposing its functionality via a FastAPI endpoint and deploying it on GCP using Docker. Software Engineer Intern Keri Software Solution 12 District, Ho Chi Minh City

March 2024 - April 2024

Contributed to the development of scalable back-end services using Node.js to support core application features.

Designed and implemented the database schema for a web application, ensuring data integrity and optimized query performance.

Performed rigorous API testing using Postman to validate functionality, identify bugs, and guarantee service reliability prior to deployment.

Project: Đồng Tiến English Center Website

Languages: Vietnamese (Native), English (IELTS 6.5 Equivalent) LÊ DĨ HÀO

LLM & Transformers: BERT, LegalBERT, T5, GPT, Llama 3.3, PhoBert, VinaLlama, LangChain, LangGraph, RAG.

Core NLP Tasks: Semantic Search, Question-Answering (Q&A), Text Summarization, Text Classification, Named Entity Recognition (NER), Text Cleaning & Segmentation. Model Fine-Tuning: PEFT, LoRa, QLoRa, SFT, DPO.

Speech AI: Whisper, ElevenLabs, Silero-VAD.

Computer Vision: YOLO, Diffusion Models, GAN.

NLP & AI Models:

September 2020 - December 2024

Voice Estate Assistant - AI-Powered Real Estate Advisor Bachelor of Information Technology Can Tho University EDUCATION

TECHNICAL SKILL

Data Engineering:

Data Processing & ETL: Pandas, EDA, Multi-Processing/Threading, Ray. Vector Databases: Pinecone, Qdrant, ChromaDB.

Web Crawling: Scrapy, Selenium, Crawl4AI.

Relational Databases: PostgreSQL, MySQL, MariaDB

Deployment & MLOps:

API & Web Frameworks: FastAPI, Flask, Gradio, Streamlit, Groq API, FastRTC. Containerization & Orchestration: Docker, Docker-Compose, Kubernetes Model Serving & Optimization: Triton Inference Server, TensorRT, vLLM, Modal, TorchServe, Quantization, Multi-GPU.

CI/CD & Version Control: GitHub Actions, Gitlab, Git. Monitoring & Observability: Prometheus, Grafana, ELK Stack. Communication Protocols: MCP, A2A

Cloud & Infrastructure: AWS (EC2, S3, Boto3), On-Premise Deployment KEY PROJECTS

Github: https://github.com/Haole3435/AI_Estate_Consulting.git Developed a voice-first intelligent assistant for the real estate domain, providing users with consultancy on property purchasing, including budget planning, location scouting, and legal advice, all through a natural Vietnamese voice interface.

Architected and implemented an end-to-end system integrating Llama 3.3 70B (via Groq API) for complex reasoning, Whisper for accurate speech-to-text, and ElevenLabs for natural text-to-speech. Constructed the RAG knowledge base by sourcing and processing the 'Vietnamese Legal' dataset from Kaggle, providing a rich foundation of legal documents for the assistant. Leveraged RAG techniques to augment the LLM with this specialized knowledge base, ensuring accurate and context-aware responses to legal and real estate queries. Impact: This project showcases the ability to build sophisticated, multi-modal AI applications for specialized domains, demonstrating expertise in system integration and applying state-of-the-art voice and language models to solve real-world problems.

RAG-Based Q&A System for Document Analysis

Github: https://github.com/Haole3435/chat-pdf.git

Engineered a sophisticated RAG (Retrieval-Augmented Generation) system enabling users to query information from PDF documents using natural language. Utilized LangChain and OpenAI for advanced reasoning and integrated Pinecone for scalable semantic search, creating a powerful question-answering tool. Impact:This project proves the capability to build intelligent search systems that understand user queries and retrieve precise information from large document corpora.

Contact this candidate