At Inventeron Technologies, I developed and optimized machine learning models for classification and prediction, improving model accuracy by 15%. I also used exploratory data analysis, preprocessing, and feature engineering to reduce inconsistencies across 10,000+ records.
I built end-to-end ML pipelines and data visualizations using Matplotlib, reducing data processing time by 25%. That work enabled data-driven insights.
On GrantLegally, I engineered an AI-powered legal intelligence platform that combines RAG, vector search, and LLM-based document understanding. It includes semantic document retrieval, AI legal chat, and bail recommendation workflows.
On FinAnalytica, I developed a financial analytics platform that automates statement analysis and reduced manual valuation effort by 60%. I also published research on visual speech recognition using a CNN with Time-Distributed layers trained on the MIRACL-VC1 dataset.
