Lavish Gangwani
Software Engineer + Developer, AI Engineer, System Designer, ML Engineer, and Researcher
AI-powered HRMS System
A smart HR assistant built with LLMs to handle leave queries, document search, and onboarding. Integrated with FastAPI, MongoDB, Redis, and Azure AI for seamless workflow automation and real-time employee support.
Multimodal Marine Assistant
An AI-powered assistant for marine engineers using OCR-parsed documents, real-time image uploads, and a domain-specific RAG system trained on proprietary marine data and a context-aware model with memory for accurate and multimodal responses.
AI-Powered Proctoring & Cheating Detection System
A comprehensive real-time and post-exam vigilance engine leveraging AI and Computer Vision. The system processes webcam/CCTV feeds using YOLOv8 for object detection (e.g., phones, books), pose and eye tracking for behavioral cues, and dynamically generates PDF-based incident reports. Enabled robust exam monitoring through live alerts and post-exam audit tools, ensuring fairness and minimizing human supervision overhead.
Hi there
I'm Lavish Gangwani, an AI Engineer based in Mumbai, focused on building intelligent systems that automate decisions, reveal insights, and create tangible value.
My journey began in the world of commerce, where I became deeply curious about how data shapes decisions. That curiosity soon turned into code—pushing me to explore machine learning, large language models, and end-to-end AI solutions.
Over time, As I deepened my understanding of AI systems, I gravitated toward solving problems end-to-end—going beyond just building models to designing complete, production-ready pipelines. I learned how to move from experimentation in notebooks to building scalable APIs, designing robust data flows, and integrating models into real applications. This process taught me not just the science of machine learning, but the engineering discipline behind reliability, iteration, and maintainability.
Today, I focus on designing AI tools that are not just experimental, but dependable—systems that do their job quietly and well.
Send me a message