Information Technology Ai Engineer

Tuan-Vinh La

+84-943-***-*** **************@*****.***

ï tuan-vinh-la Tuan-Vinh La

Artificial Intelligence Engineer

OBJECTIVE

I am an AI Engineer with 3.5 years of experience, currently I’m working at Bosch Global Software Vietnam, specializing in Computer Vision, NLP, Graph Neural Networks, and Generative AI. I apply advanced AI techniques to solve complex business and engineering problems, from design to real-world deployment. EDUCATION

• Vietnam National University - University of Information Technology 02/2022 - 12/2024 Master - Computer Science Ho Chi Minh City, Vietnam

• Vietnam National University - University of Information Technology 09/2016 - 03/2021 Bachelor - Computer Engineering Ho Chi Minh City, Vietnam EXPERIENCE

Bosch Global Software Technologies Company Limited, Vietnam AI Engineer, Sep 2022 – Present

• Multi-Angle OCR: Extract text from design documents and technical drawings with varying text angles for manufacturing and quality assurance.

Gathered, cleaned, generated, and labeled data.

Utilized YOLO to detect oriented bounding box characters/text in different angles.

Performed text recognition on rotated bounding boxes.

Developed a website for users to upload documents and perform extraction.

Technologies: Python, PyTorch, OpenCV, PaddleOCR, YOLO OBB, Streamlit.

• Food Classification for Oven: Real-time food recognition system for smart ovens to automatically adjust cooking settings.

Gathered and labeled data.

Built a lightweight, real-time model with high accuracy.

Visualized heatmaps to evaluate overfitting.

Deployed margin loss to reduce overfitting.

Technologies: Python, TensorFlow, PyTorch, OpenCV.

• Invoice Extraction: Automate invoice data extraction for accounting and ERP integration.

Utilized a lightweight model to extract text.

Deployed logic to handle inconsistent angles.

Technologies: Python, TensorFlow, PyTorch, OpenCV.

• Internal OCR Tool: Extract text from scanned Excel files and reconstruct tables.

Checked input data quality using gradient-based image processing.

Applied Hough line detection to extract tables.

Used template matching to detect icons beside text.

Applied PaddleOCR to recognize text.

Technologies: Python, PyTorch, OpenCV, PaddleOCR.

• Internal Policy QA Chatbot (Generative AI Project): Internal chatbot enabling employees to query internal policies from unstructured documents.

Collaborated with business stakeholders to define requirements.

Designed conversation flows and system architecture.

Built data pipelines to process multi-format documents (PDF, DOC, XLSX, PPTX).

Optimized retrieval and reduced response times.

Developed automated web crawlers to update content.

Built a Text-to-SQL module integrating LLM-based query translation and dynamic SQL execution.

Developed RESTful APIs for integration.

Technologies: Python, LlamaIndex, OpenAI, Hugging Face, Selenium, BeautifulSoup, FastAPI, Azure, MySQL, Docker.

• JD and CV Matching: AI system matching job descriptions to candidate CVs.

Defined use cases and system evaluation strategy.

Created pipelines for data cleaning and ingestion into vector databases.

Enhanced prompt engineering to improve matching performance.

Developed APIs for integration.

Technologies: Python, LlamaIndex, OpenAI, Hugging Face, FastAPI, HTML, SCSS, MySQL, Docker.

• Icon Detection and Classification: Detect and classify icons in product documentation and UI screenshots to support cataloging and validation workflows.

Developed image processing pipelines to detect icon bounding boxes.

Implemented template matching for icon classification against the known database.

Designed logic to flag and recognize unknown icons not present in the reference dataset.

Technologies: Python, OpenCV.

• Face Verification for Transactions: Face verification system for CIMB Bank to secure transactions.

Developed modules for face detection and recognition to authenticate users.

Implemented face liveness detection to prevent spoofing attacks.

Built face occlusion detection to ensure image quality.

Integrated deepfake detection to enhance security against synthetic media.

Technologies: Python, PyTorch, OpenCV, Tensorflow.

• Front-End Development: Customer management system for internal staff.

Developed Angular application for managing customer information.

Supported UI/UX logic improvements.

Wrote unit tests achieving test coverage >80%.

Technology: Angular.

Graduate Research Assistant, Vietnam Researcher, May 2021 – May 2025

• Conducted advanced research in multimedia computing, computer vision, NLP, and graph neural networks.

• Designed and implemented novel methods addressing limitations of existing approaches.

• Optimized algorithms, implemented robust code, and performed extensive experimental evaluations.

• Lead author on multiple peer-reviewed research publications.

• Mentored junior researchers and contributed to conferences and grant proposals.

• Research Domains: Multimedia Computing, Computer Vision, NLP, Graph Neural Networks, Knowledge Graphs, Vision-Language Models.

Emage Development, Vietnam AI Engineer, Mar 2021 – Jun 2022

• Defect Detection System: AI-powered defect detection for production lines.

Designed a system to detect product defects in manufacturing.

Developed object detection algorithms for real-time identification.

Technologies: Python, PyTorch, TensorFlow, TensorRT.

• Robotics and Sales Chatbot: Robotics solutions and customer consultation chatbot.

Developed robotic features including object detection and question answering.

Designed chatbot system for sales support.

Technologies: Python, PyTorch, Word Embedding, BM25, TF-IDF, Semantic Search. PUBLICATIONS

• KGAlign: Joint Semantic-Structural Knowledge Encoding for Multimodal Fake News Detection.

• Knowledge Graphs and Fine-Grained Visual Features: A Potent Duo Against Cheapfakes (Under Review).

• Leveraging knowledge graphs for cheapfakes detection: Beyond dataset evaluation 2023 IEEE International Conference on Multimedia and Expo Workshops (ICMEW) .

• Leverage Boosting and Transformer on Text-Image Matching for Cheap Fakes Detection. In Algorithms, 2022.

• A Combination of Visual-Semantic Reasoning and Text Entailment-based Boosting Algorithm for Cheapfake Detection. In ACM Multimedia 2022.

• A Textual-Visual-Entailment-based Unsupervised Algorithm for Cheapfake Detection. In ACM Multimedia 2022.

• Multimodal Cheapfakes Detection by Utilizing Image Captioning for Global Context. Proceedings of the 3rd ACM Workshop on Intelligent Cross-Data Analysis and Retrieval. 2022.

• Improving the Awareness of Sustainable Smart Cities by Analyzing Lifelog Images and IoT Air Pollution Data. IEEE International Conference on Big Data (Big Data). IEEE, 2021. SKILLS

• Programming Languages: Python, C++

• Frameworks & Tools: TensorFlow, PyTorch, Hugging Face, LlamaIndex, Git, FastAPI, Angular, MongoDB

• DevOps & Version Control: Git, GitHub

• Specializations: Image Processing, Computer Vision, Natural Language Processing, Knowledge Graphs, Graph Neural Networks, Generative AI, Multimedia Computing

• Foundation Knowledge: Linear Algebra, Statistics, Mathematics for Machine Learning, Mathematics for Deep Learning

HONORS AND AWARDS

• Grand Challenge Champion in Cheapfakes Detection Task 2 — IEEE International Conference on Multimedia & Expo 2023.

• Certificate of Commendation for Research Excellence — Awarded by the University of Information Technology, Vietnam National University (2022).

Contact this candidate