MANJU A R

I’m a Computer Science and Engineering graduate with a strong foundation in Generative AI and Machine Learning. I build AI-powered applications end to end, turning models into usable products that solve real-world problems.

My technical strengths include Python-based development with PyTorch, Hugging Face, and fine-tuning workflows for LLMs. I’ve applied prompt engineering and RAG approaches, and I’m comfortable optimizing inference pipelines for performance and reliability.

I’ve delivered projects like a real-time Text-to-Image Generator using Stable Diffusion with a Gradio interface, plus an AI-powered Forensic Face Sketch Construction and Recognition system. In that work, I used MTCNN detection, FaceNet embeddings, and developed Flask REST APIs for image processing and face matching.

Through internships and data projects, I’ve also strengthened my data and analytics capabilities—cleaning datasets, running EDA, doing feature engineering, and building predictive models. For example, in a traffic simulation project using SUMO and Q-Learning, I reduced average vehicle wait time by 12%.