Open to opportunities

Hemanth S R User

@hemanthsruser

Message

I build machine learning models and annotate data for AI products.

India

Message

At Samsung Seed Lab, I analyze speech, image, text, video, and segmentation projects, supporting quality assurance, data verification, and Python automation. I label object-detection data using bounding boxes and polygons, along with text and audio data.

I've built speech emotion recognition with CNN-LSTM networks, an 82.53% accurate handwriting recognition system, surgery mortality prediction models in PyTorch, and a cosine-similarity movie recommender. Previously, at Varcons Technologies, I worked on plant tumor disease detection using CNN, RNN, and GAN architectures.

Experience

Work history, roles, and key accomplishments

Current

AI Trainer

Current

Samsung Seed Lab

Nov 2024 - Present (1 year 9 months)

Analyzing projects such as Speech Data Collection, Image Data Annotation, Semantic Text Segmentation, Unified Clipper Annotation, Metric Video Annotation, Semantic Segmentation, Quality Assurance and Validation, Data Verification and basic automation scripts with Python. Data Labelling for object detection, bounding boxes, polygon annotation, text and audio labelling.

Python Data Annotation Quality Assurance Object Detection Bounding Boxes Polygon Annotation Audio Labeling

Intern

Varcons Technologies Pvt Ltd

Oct 2023 - Feb 2024 (4 months)

Part of developing a project to detect plant tumor diseases using machine learning algorithms, achieving a result rate using architectures such as Convolution Neural Networks, Recurring Neural Networks and General Adversarial Networks.

Machine Learning Convolutional Neural Networks Recurrent Neural Networks

Student Developer

Hackathon: HACKOVERFLOW 5.0

Mar 2022 - Mar 2022 (0 months)

Participated in collaborative programming project with a team of 2 people. The project was a Railway Ticket Booking in a working Prototype of a website that could let an individual select his/her own seats while booking train tickets. The development of the project was done using HTML, CSS and Javascript.

HTML CSS JavaScript