Omkar Rumane

I’m an Associate Software Engineer with hands-on experience building AI-driven backend services, predictive models, and RAG-based LLM systems. I focus on developing end-to-end AI pipelines using Python and FastAPI, from data preprocessing and model optimization to deploying solutions into production environments.

At Altysys, I engineered high-performance predictive models using XGBoost and CatBoost to tackle binary classification and regression problems. I also implemented probability calibration and optimization in Python to improve the reliability and correctness of ML outputs, and I built scalable full-stack applications integrating a Next.js frontend with a FastAPI backend.

Earlier, I worked as a Research Intern at Energy Research Institute @ NTU (ERI@N), where I synthesized research from 4+ academic papers and developed test case scenarios in the CARLA simulator. I also contributed to AI projects like an AI Legal Assistant multi-agent RAG pipeline (FastAPI, LangChain, FAISS, Groq/Qwen) and achieved strong results with models such as a Random Forest car evaluation classifier (94% accuracy), alongside publishing a technical article and completing an IBM Business Analytics certification.