Ritik Raj Email & Phone Number

Hyderabad, In

Hyderabad, Telangana, India

Packet Fraud Detection- Led a proof of concept (POC) project for verifying integrity and detecting fraud, followed by OCR-based extraction of unique packet IDs for validation.- Developed a classification model to verify if packets were opened or remained intact, utilizing YOLO for object detection to identify packet conditions, detect signs of tampering or fraud, and detect text region areas.- Integrated an OCR system to extract and validate unique packet IDs from various packet surfaces, collecting and annotating a diverse dataset of packet images to train the models.- Optimized model performance and inference speed through techniques such as model quantization.- Skills: PyTorch · OpenCV · Python . Efficient Net, YOLOv8

Hyderabad, Telangana, India

Automated Label Validation- Implemented deep learning models using PyTorch, Yolov5, OpenCV, Spacy, Sentence Transformers, and Sklearn to improve accuracy and reduce costs.- Built a proactive learning framework for object detection models and implemented retraining pipeline and drift monitoring on unstructured image datasets.- Collaborated with cross-functional teams, including software developers, and annotators, to design and implement the platform.- Conducted user testing and gathered feedback to continuously improve the platform. As a result, the platform achieved a 95% accuracy rate and significantly reduced the time required for validating product labels.- Expanded the platform for multiple product portfolios and packaging types in different geolocations. Successfully led the development of a computer vision platform that automated label validation tasks for a large FMCG company. Skills: PyTorch · OpenCV · TensorFlow · Python (Programming Language) · Natural Language Processing (NLP) · Computer VisionEmerging Bussiness Opportunity :- Designed a framework to identify emerging trends from claims, ingredients, and flavors of products in the pet consumables category and forecasted their potential growth using Google Trends and Sprinklr data for a large FMCG company.- Created leading indicators based on a novel approach using the Dynamic Time Wrapping algorithm.- Skills: Time Series, Python (Programming Language)

New Delhi, Delhi, India

Flood Risk Assessment :- Involved in extracting GIS data from multiple sources which include raster data, and so on. - Implemented multiple GIS algorithms on raster data and incorporated parallel processing and NUMBA to speed up the data extraction process- Involved in data cleaning, data analysis process.- Worked on creating multiple classification models and have done a comparative analysis.

New Delhi, Delhi, India

Involved in Data Preprocessing Techniques for making the data useful for creating MachineLearning modelsWorked on building web scrapperWriting shell script to automate the tedious task.Worked on databases such as MongoDB and PostgreSQLData preprocessing for Geographic Information System(raster data, Geojson data)

M.Tech, Computer Science

M.Tech Research Project, Deep Learning Based Liveliness Detection And Voice Authentication

Bachelor Of Technology, Computer Science