Jeevan K.

Senior Computer Vision Research Engineer & AI Specialist with extensive experience in designing and deploying advanced AI models for industrial and biomedical applications. OCR Pipeline for Automated Optical Inspection Developed a high-accuracy OCR system to extract textual information from electronic component images, achieving 95% accuracy and significantly improving manufacturing inspection workflows. Custom Segmentation and Defect Detection Models. Engineered YOLO-based models that reduced segmentation runtime from 35 ms to 3 ms, enhancing real-time defect detection in surface mount technology. Natural Language Processing for Manufacturing Data Created NLP methods to accurately extract Manufacturer Part Numbers (MPN) from Approved Vendor Lists (AVL) and catalog descriptions, automating data processing pipelines. PUResNet (V1.0 & V2.0): Protein-Ligand Binding Site Prediction Designed and implemented deep residual UNet architectures utilizing sparse Minkowski tensors for efficient large-scale protein structure analysis, advancing computational drug discovery. Post-Translational Modification (PTM) Site Prediction via Protein Language Models Fine-tuned GPT-based protein language models using prompt-based unsupervised learning to predict PTM sites and analyze mutation effects near functional regions. NanoBinder: Nanobody-Antigen Binding Prediction Tool Developed an interpretable machine learning pipeline integrating Rosetta energy scores, fully automated data extraction, and web deployment for robust nanobody binding affinity prediction. Custom Autoencoder Models for Image Clarity Assessment Designed tailored autoencoder architectures to evaluate and enhance image clarity of electronic components in manufacturing inspection systems. Web Application Deployment with Django Built scalable backend services to host AI models, enabling public accessibility and integration with client workflows. Technical Skills & Tools: Python, PyTorch, TensorFlow, Huggingface Transformers, YOLO, OpenCV, Docker, Slurm HPC, Git, Pymol, AutoDock Vina, Gromacs, Schrodinger, openMM. Academic Excellence: Ph.D. candidate in Integrated Energy AI with multiple publications in top-tier journals such as Nature Communications, Journal of Cheminformatics, and Computational and Structural Biotechnology Journal.