Diya Manoj

I’m an AI & Data Science undergraduate focused on turning messy, real-world data into working AI systems. I enjoy rapidly prototyping and validating end-to-end pipelines that reliably support practical decision-making.

In my projects, I’ve built an end-to-end CCTV Anomaly Detection System using CNN autoencoders, including preprocessing pipelines and reconstruction-error–based anomaly scoring with configurable thresholds. I also delivered the full workflow—from data ingestion and model training to evaluation and Streamlit visualization.

Right now, I’m developing Unify, an AI-powered intelligent preprocessing system that uses LLM-driven workflows to transform unstructured data into structured, ML-ready formats. I’m building modular, interpretable data processing pipelines using controlled prompt engineering and retrieval-style context handling to reduce manual preparation and improve reliability of downstream AI systems.

Professionally, I’ve worked as a Data Analyst Intern, cleaning and validating real-world datasets, building Power BI dashboards for KPI monitoring, and translating stakeholder requirements into actionable insights. I’ve also completed an AI Intern role where I scoped feasible chatbot solutions under real-world constraints, and I contributed to product innovation at IIT Bombay’s Invention Factory through design thinking and prototype development.