As an AI/ML Developer and Backend Engineer, I have worked on various projects, including the development of legalproai.in and other applications. My experience emphasizes my skills in machine learning, computer vision, and backend development.Key Projects: 1. LegalProAI: Developed an AI-driven platform for legal services.2. AI-Powered Chatbot: Created a chatbot for law enforcement to analyze crime data.3. Object Detection Models: Implemented models for object detection and segmentation in collaboration with IFB Industries.4. Person Re-Identification System: Developed a system using face recognition technology to improve tracking accuracy.5. Vehicle Re-Identification Projects: Contributed to projects that reduced identification errors in vehicle tracking.6. Computer Vision Security Systems: Established systems for security applications, optimizing performance and reliability.SkillsProgramming Languages: Python, Java, C, C++Deep Learning Frameworks: TensorFlow, PyTorch, KerasWeb Development: FastAPI, Flask, DjangoCloud Platforms: AWS, Azure AIData Visualization: Matplotlib, SeabornDevOps Tools: Docker, GitMy commitment to leveraging cutting-edge technologies has enabled me to deliver impactful solutions that meet client needs.

Collaborated on a team project detecting item quantities in a refrigerator. Upgraded data pre-processing and selected an optimal Deep CNN algorithm, increasing accuracy from 67% to 89%. Developed automated data collection scripts, boosting productivity by 50% and saving company resources. Built and deployed object detection, segmentation, and key-pointsdetection models to production.• Collaborated extensively with the R&D team, Developed, optimized, and deployed advanced models for object detection, segmentation, and key-point detection, resulting in a 25% improvement in accuracy compared to the previous model and a 30% reduction in false positives.• Leveraged state-of-the-art models such as YOLO, Detectron2, and SAM, along with the OpenCV library, achieving an average precision score of 0.92 on the custom dataset. • Employed image processing scripts to enhance data pre-processing, resulting in a remarkable accuracy boost from 67% to 92%.• Automated data collection procedures through the development of custom scripts, leading to a 50% reduction in data preparation time and freeing up valuable resources for other critical tasks.Technologies & Topics: Pytorch, AWS, Yolo, Instance Segmentation, Key-point Detection

Contributed to the development and enhancement of a real-time video analytics system, incorporating advanced computer vision algorithms for accurate object recognition, tracking, and behavior analysis.1. Person Re-Identification For Ticket Grading:• Developed and executed a person re-identification system leveraging face recognition technology to verify consistency of individuals across their entire journey, overcoming tracking ID discrepancies.• Worked closely on person tracking use-case, utilizing public REID datasets such as Market1501, DukeMTMC-reID, and CUHKКОЗ.• Implemented repositories for person re-Identification, using techniques like Domain Adaptive Person ReID on custom dataset2. Computer Vision Projects:• Collaborated with the team on vehicle ReID projects, including tasks such as writing data transformation scripts and model conversion scripts. • Achieved success in developing, testing, and deploying projects focused on object detection and object tracking.Technologies & Topics: Pytorch, Docker, TensorRT, Triton, AWS, Person-ReId, Vehicle-ReID, Yolo, Object tracking

• Actively participated as a member of a 10-engineer team engaged in researching the company's Computer Vision and ARbased product.• Led the design and implementation of more than 5 3D object pose detection algorithms, meticulously evaluating theirviability and effectiveness.• Prepared and presented a thorough 10-page written report, outlining a comprehensive work plan encompassing all aspects of the product.

• Authored and published a series of 12 impactful technical blogs on a website, attracting a consistent traffic of 650,000 views.• Successfully reached a broad audience of 59.3 million individuals by sharing insightful blogs on cutting-edge technologies such as Data Science, Computer Vision, Deep Learning, and Artificial Intelligence.• Assumed the responsibility of peer-reviewing blog articles, meticulously reviewing and providing constructive feedback on 15 blog posts to enhance their overall quality and impact.• Demonstrated expertise in writing engaging blogs on diverse technologies, including C++, Python, and OpenCV, while showcasing proficiency in Machine Learning, TensorFlow, Data Science, Seaborn, and the Python programming language.

1. My role was to develop, test the python programs for them which I have made.2. The code snippets are used by the developers for building real-life applications.3. The project was fully about OpenCV, I worked on various projects with the outcome of Computer Vision.