Om Khairate

Computer Vision & Machine Learning Researcher

Saarbrucken, DE.

About

Highly motivated Computer Vision and Machine Learning Researcher with a Master of Science background, specializing in advanced deep learning architectures and their real-world applications. Proven ability to conduct impactful research, optimize complex systems for performance and robustness, and contribute to peer-reviewed publications. Seeking to leverage expertise in cutting-edge R&D roles within academic or industrial research environments.

Work

August-Wilhelm Scheer Institute for digitized products and processes gGmbH

|

Research Assistant

Saarbrucken, Saarland, Germany

→

Summary

Conducting advanced research and implementing state-of-the-art deep learning architectures and methods to drive innovation in digital product and process development.

Highlights

Researched and implemented state-of-the-art (SOTA) deep learning architectures, contributing to advancements in digital product and process optimization.

Developed novel methodologies for data analysis and model training, enhancing efficiency and accuracy of research outcomes.

Collaborated with interdisciplinary teams to integrate deep learning solutions into complex systems, demonstrating strong problem-solving skills.

NextLeap Aeronautics Pvt. Ltd.

|

Computer Vision Intern

Pune, Maharashtra, India

→

Summary

Engineered a deep-learning helipad detector with Euler angle estimation and closed-loop control, validated on a physical drone for extreme-environment testing.

Highlights

Developed a deep-learning helipad detector, integrating Euler angle estimation and closed-loop control for autonomous drone landing systems.

Validated system performance on a physical drone, demonstrating robust functionality and precision in extreme environmental conditions.

Designed and implemented custom data acquisition tools, supporting comprehensive testing and refinement of the helipad detection model.

Utilized SSDMobilenet model for helipad detection and Mavlink messages via Mavsdk platform to achieve vision-based landing capabilities.

Education

Universität des Saarlandes

Saarbrucken, Saarland, Germany
→
Master of Science Computer Science

Vishwakarma Institute of Technology

Pune, Maharashtra, India
→
Bachelor of Technology Instrumentation and Control Engineering Grade: 9.17

Nagesh Karajagi Orchid School

Solapur, Maharashtra, India
→
Higher Secondary School Certificate Science

St. Joseph High School

Solapur, Maharashtra, India
→
Secondary School Certificate General Studies

Awards

Second Best Hack Award

PickHacks '21 by Missouri University of Science and Technology

Awarded for outstanding performance in a hackathon, demonstrating innovative problem-solving and technical skills. Project: rap-god-42mnyv.

Best Hardware Hack Award

RowdyHacks '21 by University of Texas San Antonio

Recognized for developing the most innovative hardware solution in a competitive hackathon. Project: crashbuds.

Second Best Hack Award

HighField Hacks '21 by SouthHampton University

Received an award for exceptional project development and execution in a hackathon setting. Project: gym-buddy-rnt9yp.

Best Safety Hack Award

Machacks '21 by McMaster University

Honored for creating the most impactful safety-focused solution during a hackathon. Project: drive-buddy.

Publications

A Study on the Robustness of Object Detectors in Aqua-Farming

SCITEPress

Analyzed transfer-learning robustness of object detectors for fish detection, achieving 0.92 mAP in-domain (YOLOv8l) and quantifying out-of-domain false positives; evaluated retraining and bias-only adaptation strategies.

Skills

Programming

C++, Python, Metal Programming.

Computer Vision

Image Processing, Image Analysis, Image Segmentation, Computer Graphics.

Machine Learning / Deep Learning

Deep Learning, Machine Learning, Data Preprocessing.

Systems / Tools

Linux.

Projects

Analysing Automated Monitoring of Industrial Water Treatment Plants: From IoT Sensing to Computer Vision Approaches

→

Summary

Surveyed and implemented vision-based pipelines for industrial wastewater monitoring. Focused on anomaly detection, aeration analysis, and camera-based water-level estimation under real-world constraints.

IoT based Shower Monitoring System (Patent ID: 202221009441)

→

Summary

Developed an IoT-based smart shower system with automated water cutoff, environmental monitoring, and an emergency override function to significantly reduce water consumption.

Drowsiness Detection System

→

Summary

Developed an early-warning system for drowsy driving using machine learning algorithms to detect open and closed eyes, enhancing road safety and preventing accidents.

License Plate Recognition System using YOLOv4

→

Summary

Leveraged YOLOv4 architecture to train a system for detecting license plates with an optical character recognition algorithm, outputting the license plate number.