Cheonan, South Korea · IEEE Peer Reviewer
Press ⌘K for quick nav
AI Research Engineer
Cheonan, South Korea 🇰🇷
Featured
Annotation-efficient framework for simultaneous instance-level teeth segmentation and missing-tooth detection in panoramic radiographs. Submitted to Computers in Biology and Medicine.
Real-time computer vision pipeline for industrial defect detection, integrated into a C#/WPF machine UI. Deployed in production with Docker and FastAPI backend.
About
I build deep learning systems for medical imaging — focusing on dental panoramic radiographs and liver MRI — to support clinical decision-making. My work spans model design, training, and deployment, with publications in IEEE, Knowledge-Based Systems, and Computers in Biology and Medicine.
Beyond medical AI, I develop applied AI systems for real-world environments: LLM pipelines, multi-agent research assistants, and computer vision integrated into industrial UIs. I work across the full stack — from research prototype to production deployment.
Actively seeking opportunities in Medical AI, LLM Engineering, and Computer Vision — research and industry alike.
Skills
Experience
Led end-to-end low-data industrial defect detection pipeline — dataset strategy, training, evaluation, and deployment into operator-facing C#/WPF inspection workflows. Developed adaptive methods robust to distribution shift; implemented uncertainty-aware inference for ambiguous case flagging.
Research on annotation-efficient and robust deep learning for medical image segmentation and localisation. Systematic ablation studies on model generalisation under noisy and imbalanced data. Co-authored multiple peer-reviewed publications.
Research on industrial automation communication systems (Ethernet/IP). First-author publication at IEEE ICCAS 2022.
Led a technical team from concept to deployment on engineering solutions, including AI integration and systems design.
Projects
Research Projects
Annotation-efficient deep learning for instance-level teeth segmentation and missing-tooth detection in panoramic radiographs. Under review at Computers in Biology and Medicine.
Hybrid CNN–Swin Transformer model for automated cirrhotic liver segmentation on MRI. Published in IEEE Access.
Autonomous pipeline (Planner · Researcher · Writer · Editor) integrating GPT-4, arXiv, Tavily, Wikipedia. Dockerized CLI generates Markdown/HTML reports.
Applied AI & Deployments
Autonomous pipeline (Planner · Researcher · Writer · Editor) integrating GPT-4, arXiv, Tavily, and Wikipedia. Dockerized CLI that generates structured Markdown/HTML reports.
LLM-powered Slackbot for content ideation and keyword extraction. Deployed live on Render; integrates Slack Events API and OAuth.
Real-time CV defect detection integrated into a C#/WPF machine UI with ONNX Runtime and FastAPI backend services.
Publications
Journal Articles
Conference Papers
Patents & Poster
Patent Issuer: Flexigo Co., Ltd. (Korean) · Shaily Bajpai and Lee Ki Yong
Patent Issuer: Chosun University Industry-Academia Cooperation Foundation (Korean) · Shaily Bajpai and Bumshik Lee
36th Workshop on Image Processing and Understanding (IPIU 2024) · Shaily Bajpai and Bumshik Lee
Certifications & Awards
Awarded for filing a patent on Deep Learning-Based Vision Inspection Device and Method in an industrial production environment.
Awarded for the presentation of Advanced Teeth Instance Segmentation Using Contextual Spatial Attention and Deeper Mask Head.
Consecutive research scholarship awarded for all four semesters of the MSc programme based on academic and research performance.
Certification in Agentic AI systems design and implementation.
Certification covering large language model architectures, fine-tuning, and production deployment patterns.
Containerisation and cloud deployment on Microsoft Azure with Docker.
Education
Major: Deep Learning / Artificial Intelligence · GPA: 4.31 / 4.5
GPA: 7.23 / 10 · First Division
Contact
Get in touch
Open to research collaborations, PhD opportunities, and industry roles.
For fastest response, email is best.
I typically reply within 24–48 hours.
shailybajpai.01@gmail.com