Tuan D.

I am a Senior AI Engineer with 5+ years of experience building production-grade computer vision and LLM-powered systems for logistics and media companies. I specialize in solving complex document processing challenges and building intelligent systems that work reliably in production environments. What I Solve For You: - High-Precision OCR: build custom pipelines for complex layouts. At CyberLogitec, I improved container number recognition accuracy from 91% to 98.1% by optimizing post-processing logic. - Document digitization and key-value extraction from PDFs, shipping documents, and invoices - Object detection, classifiication and segmentation with automated dataset optimization - Image retrieval systems for enterprise applications - Advanced RAG pipelines with hybrid retrieval (keyword + semantic search) and re-ranking - Robust document ingestion pipelines for PDF, DOCX, XLSX, and PPTX files - Automated pipelines that turn your messy, scanned, and rotated PDFs into clean, validated Excel/JSON data. My Tech Stack: - OCR: PaddleOCR, MMOCR, EasyOCR, Tesseract, TrOCR - Detection: YOLO (v5-v26) - Deployment: Docker, Flask/FastAPI, Azure MLOps - LLM Frameworks: LangChain, LlamaIndex - LLM Models: GPT-4, Gemini (Vertex AI), Open Source (Llama 3, Mistral) - Vector DBs: Chroma, FAISS, Pinecone - Data Extraction: PyMuPDF, pdfplumber (for digital PDFs), Google Document AI (for enterprise scale).