kavitha pathmanathan

I am a Machine Learning Engineer with 3.8+ years of experience applying generative AI, computer vision, NLP and data science to production problems. I build and deploy scalable ML solutions, from model training to Dockerized FastAPI inference and cloud deployment.

At Orel IT I engineered RAG chatbots, profiling and rating systems using LlamaIndex and GPT models, and created a generative workflow agent for supermarket planogram automation that enforces complex hierarchies.

I developed facial verification and multi-person tracking systems using MTCNN, VGGFace2 embeddings and DeepSORT, fine-tuned a Vision Transformer (DeiT) on 8K+ products, and implemented Siamese, UNet and LSTM models for product similarity, segmentation and activity classification.

I am pursuing an MSc in Data Science & Artificial Intelligence and have published work on deep learning for music classification and segmentation-model analysis; I focus on delivering research-grounded, production-ready ML systems that solve business problems.