Rumit Pathare

About Rumit Pathare

I build real-world AI systems that solve industrial problems.I’m a Machine Learning Engineer working at the intersection of Computer Vision, Reinforcement Learning, and production-grade ML architecture for manufacturing and engineering use cases.At WonderBiz (Schneider Electric client projects), I’ve architected reusable AutoML platforms, CNN-based vision systems for interpreting engineering diagrams, and a reinforcement learning simulator that became the foundation of an international AI patent for water optimisation filed across 6 countries.What makes my work different is my background.Before ML, I worked in production engineering, managed shop-floor operations, and even worked as a 3D visualisation artist. This gives me a systems-level understanding of how real processes, visuals, and constraints work — which directly translates into how I design ML solutions today.Key highlights:• Built AutoML infrastructure reused across 3 enterprise POCs that converted into paid projects • Designed a U-Net + MobileNetV2 system achieving 93% accuracy for industrial diagram segmentation • Engineered an OpenAI Gym RL simulator for water optimisation (International patent co-inventor) • Mentor to engineers and contributor to open-source CV and OCR projectsI’m particularly interested in CNN research, computer vision, and reinforcement learning applied to real-world systems.

Experience

1. Software Engineer

   WonderBiz Technologies Pvt.

   Feb 2023 — Present · IN

   Working on enterprise AI systems for Schneider Electric projects across manufacturing, computer vision, and reinforcement learning domains.• Architected a Dockerized AutoML platform (FastAPI + MSSQL) executing 8–9 algorithms, reused across 3 manufacturing POCs that converted into paid projects• Designed a CNN-based semantic segmentation system (U-Net + MobileNetV2, PyTorch) achieving 93% accuracy for interpreting P & ID engineering diagrams• Engineered a reinforcement learning simulator using OpenAI Gym for industrial water optimisation, forming the basis of an international AI patent• Built predictive maintenance and process optimisation models using sensor data and XGBoost• Developed an internal RAG system using Cohere embeddings, ChromaDB, and Streamlit for enterprise knowledge access

Education

• Coursera

  Specialization in Python for Everybody, Information Technology

• IT Vedant

  Data Science

• University of Mumbai

  Bachelor of Engineering - BE, Mechanical Engineering