Noor Uddin

Designed and deployed deep learning models (MSC-LSTM, MLP, CNN) and prompt engineering tools for anomaly detection, behaviour analysis, and colour classification. Developed NetFault_LLM for broadband fault detection and open-sourced multiple ML projects with CI/CD integration.

Fine-tuned LLMs using SFT and RLHF, conducted advanced code and data analysis, and deployed optimized models with low latency. Delivered updates on schedule, communicated insights effectively, and was promoted to Secondary PoD Lead for performance.

Reviewed GitHub projects on autoencoders, GANs, diffusion models, and GNNs, and implemented image-to-image translation using diffusion models. Also explored out-of-distribution detection and active learning techniques.

Developed and deployed an ML-powered backend system for crop disease detection using image classification, Flask API, and cloud infrastructure in 2022.

Worked as a programming instructor and developer, specializing in Python and C++, delivering structured content, developing practical software projects, conducting code analysis, and actively engaging with the team to ensure collaboration and technical alignment.

Thesis: Indoor Positioning System for Finding the Location and Predicting Missing Fingerprints Using Generative Adversarial Network (GAN).

Thesis: Sign Language Interpreter/ Translator (Signs to Voice/Text) for Mute People Using VS Studio, C#, Google Voice API, and MS Kinect Device.