SHIVANSH GUPTA

I’m a data-focused learner exploring AI and Machine Learning, with hands-on experience building LLM-based tools and data-driven applications using Python. I enjoy translating complex data problems into practical pipelines, then tightening them for real-world performance through experimentation.

In my projects, I engineered a scalable Python RSS ingestion pipeline (570+ daily articles from 26 sources) and deployed a Groq LLaMA 3.1 8B system for real-time zero-shot classification of 11 propaganda categories. I also built semantic analysis using SentenceTransformers embeddings with DBSCAN clustering and UMAP, uncovering 42 event clusters and 19 contested narratives, while containerizing FastAPI backends with Docker and CI/CD to keep daily data refresh reliable.