Madhav Murali

Developed and enhanced enterprise-scale applications for the UBSAdvantage platform using .NET (C#), Python, and TypeScript, following SDLC and Agile practices. Built Cypress-based end-to-end automation and release CI/CD orchestration, and contributed to an agentic AI platform using REST API integrations and RAG for financial insights.

Co-authored and published a research paper on a two-phase UAV approach for survivor identification in disaster-hit areas, accepted at the ICIoTCT International Conference. Designed UAV trajectory planning using Spiral Reconnaissance and TSP optimization, and developed vision/depth components with deep learning models for real-world validation.

Developed and deployed a machine learning model to recommend lenders to loan applicants based on credit and financial factors. Built and integrated a full-stack web application with React, Tailwind CSS, and REST APIs, including a React/Firebase admin dashboard for authentication and configuration workflows.

Implemented web-navigation automation for autonomous task execution using Large Language Models (LLMs) with DOM parsing and interaction pipelines. Evaluated performance on the MiniWob++ benchmark and researched in-context learning, environment grounding, and sequential decision-making for AI-driven web agents.

Activities and societies: Two-phase UAV trajectory planning combining Spiral Reconnaissance and TSP-based waypoint optimization; developed and validated path-planning and perception components using real-world DJI Tello experiments and models including YOLOv8s.

Co-authored a research paper accepted at the ICIoTCT International Conference on a two-phase UAV approach for survivor identification in disaster-hit areas.

Pursued a Bachelor's program in Computer Science at the National Institute of Technology, Calicut from 2021 to 2025.