--- title: "About" description: "CEO & Co-Founder at MicroSpace and AI Engineer with 6+ years building OCR, computer vision, NLP and generative AI systems. MSc Computer Science." url: https://around-ai.com/about author: thucpham (AI Engineer) language: en --- # About ## Hi, I'm Duc-Thuc Pham (Daniel) πŸ‘‹ **CEO & Co-Founder at MicroSpace Β· AI Engineer Β· MSc Computer Science** β€” Hanoi, Vietnam > Slow, steady, focused. I'm an AI researcher and engineer with 6+ years of hands-on experience, from OCR and computer vision to NLP and generative AI. I'm a programming enthusiast who stays curious about every corner of AI, and **around-ai** is where I document what I research and what actually works in practice. I graduated from Hanoi University of Science and Technology (HUST) in 2020, majoring in electronics and computer engineering. My daily work is applying computer vision and deep learning to build intelligent software products, from designing and training models to deploying them on mobile and embedded hardware. Before going all-in on AI, I worked as a freelance embedded hardware developer, building projects with Arduino, Raspberry Pi and Jetson Nano. Outside of my day job, I collaborate with other passionate engineers on open-source work at [nrl.ai](https://nrl.ai), including [Daisykit](https://daisykit.nrl.ai) 🍰 β€” easy AI for everyone with NCNN, with 150k+ downloads on PyPI. --- ## Experience ### CEO & Co-Founder β€” MicroSpace *Present Β· Hanoi* - Co-founded MicroSpace and lead the company as CEO. ### AI Engineer β€” Creative Force (AI Team) *Oct 2023 – Present Β· Hanoi* - Building generative AI products for e-commerce at [dreem.ai](https://dreem.ai) (team of 7). - Researching and building a solution for **generative pose transfer** for fashion models. ### AI Engineer β€” Samsung SDS Vietnam (AI Research Lab) *Jun 2022 – Oct 2023 Β· Hanoi* - Led a team of 3 to build the public demo of the team's OCR products at [ocr.copen.vn](https://ocr.copen.vn): designed an efficient, cost-effective model-serving system for many concurrent users, plus database design, storage and Active Learning. - Worked on **Key Information Extraction** and **Document Reconstruction** β€” the reconstruction system regenerates nearly 99% of administrative documents, with table reconstruction in progress. - Modified the model architecture to extract key–value pairs from long tables, and built a pipeline that auto-fills extracted information into software via RPA. ### AI Engineer β€” NTQ Solution *Apr 2021 – Jun 2022 Β· Hanoi* - Built an intelligent document system with MLOps for technical reports of a Japanese shipbuilding company (IMC): an OCR system for scanned documents (phase 1), then automated data updates, training and deployment (phase 2). Delivered excellently, with a performance bonus after each phase. - Developed a mobile facial recognition and expression detection system, including smile scoring. ### AI Engineer β€” VinAI & VinSmart *Aug 2020 – Apr 2021 Β· Hanoi* - Trained NLP models and published them on [HuggingFace/NlpHUST](https://huggingface.co/NlpHUST). - Researched image-quality enhancement for mobile cameras and helped deploy it on VinSmart phones. ### AI Engineer (part-time β†’ full-time) β€” Kaopiz, AI Department *Apr 2018 – Aug 2020 Β· Hanoi* - Built AI engines for Japanese clients: OCR, face recognition, object detection and segmentation. - Shipped AI APIs with Flask, FastAPI, Docker and AWS. - Deployed models to Jetson Nano, Raspberry Pi and mobile (TensorRT, TFLite, TorchScript, ONNX, DeepStream). ### Lab Member β€” IC Design BKIC Lab, HUST *Aug 2016 – Apr 2018* - Implemented convolution in hardware (VHDL) and got started with deep learning in Python. --- ## Education - **MSc Computer Science** β€” University of Engineering and Technology, VNU Β· 2022 – 2024 - **Telecommunications Electronics Engineering** β€” Hanoi University of Science and Technology (HUST) Β· 2015 – 2020 --- ## Honors & Awards - **2022** β€” Best Employee, Samsung SDS Vietnam - **2021** β€” Top 3, Legal Text Retrieval track, Zalo AI Challenge 2020 - **2020** β€” Best graduation project of the council, HUST - **2016** β€” Study encouragement scholarship, HUST --- ## Publications - *RR-HCL-SVM: A two-stage framework for assessing remaining thyroid tissue post-thyroidectomy in SPECT images* β€” International Journal of Imaging Systems and Technology, 2024 - *Ablation dosage recommendation for thyroid cancer treatment following thyroidectomy using machine learning* β€” Journal of Military Science and Technology, 2024 - *Utilizing DETR model on SPECT image to assess remaining thyroid tissues post-thyroidectomy* β€” IEEE Statistical Signal Processing Workshop (SSP), 2023, pp. 537–541 - *A deep learning method using SPECT images to diagnose remaining thyroid tissue post-thyroidectomy* β€” 9th NAFOSTED Conference on Information and Computer Science (NICS), 2022, pp. 93–98, [doi:10.1109/NICS56915.2022.10013322](https://doi.org/10.1109/NICS56915.2022.10013322) More on [Google Scholar](https://scholar.google.com/citations?user=wIca9RgAAAAJ&hl=en). --- ## Skills | Area | Tools | | --- | --- | | AI / ML | PyTorch, TensorFlow, Keras, Scikit-learn, OpenCV, TensorRT, ONNX, Diffusion / ComfyUI, LoRA | | Embedded & Edge | Jetson Nano, Raspberry Pi, Arduino, TFLite, VHDL | | MLOps & DevOps | AWS, Docker, Kubernetes, Kubeflow | | Back-end | Python, Django, FastAPI, Flask, REST APIs | | Other | Flutter, JavaScript | | Languages | Vietnamese, English | --- ## Get in touch - βœ‰οΈ [thuc248997@gmail.com](mailto:thuc248997@gmail.com) - πŸ’Ό [LinkedIn](https://www.linkedin.com/in/thuc-pham-807241132/) - πŸ™ [GitHub](https://github.com/thuc248997) - πŸ€— [HuggingFace](https://huggingface.co/NlpHUST) - 🐦 [X / Twitter](https://twitter.com/thuc248997)