At Cognizant, I completed PAT training in the HPC .NET healthcare domain, learning .NET technologies, SQL, and healthcare application concepts through technical training and practical exercises.
For my major project, Instruction-Tuned Generative AI Framework with PEFT and RAG for Higher Education Analytics, I implemented an instruction-tuned LLM pipeline with PEFT (LoRA) and a RAG architecture for semantic search and context-aware answers.
I used Python, Hugging Face Transformers, and embedding models in the project's training, retrieval, and inference workflows. I also built an end-to-end system for query answering, performance analytics, and personalized recommendations using Streamlit.
During my Infosys Springboard internship project, SentinelNet, I developed a machine learning-based network intrusion detection system using NSL-KDD and CICIDS datasets. Its models achieved 98% multiclass and 90%+ binary accuracy, and I designed a Streamlit interface for live checks and CSV analysis.
