At FinAnalyz Technologies Pvt. Ltd, I engineered text-classification pipelines for sentiment analysis of financial user feedback. I cleaned and engineered features from unstructured data, including bank statements, GST reports, and portfolios.
I fine-tuned tokenization, semantic chunking, and embedding pipelines, then used Python, Pandas, and machine-learning libraries to train predictive models.
I also built a RAG chatbot project for querying company documents locally. It combines LangChain document processing, Ollama embeddings and local LLMs, FAISS similarity indexing, and a Streamlit dashboard.
