P Aaditya - Portfolio

Education

Undergrad Degree:
- Course: Artificial Intelligence & Machine Learning (BE)
- Institution & University: Cambridge Institute of Technology under VTU
- Period: 2020 - 2024
Samsung Innovation Campus (SIC):
- Course: Artificial Intelligence & Machine Learning
- Period: 250 hours
- Project: Surgery Mortality Prediction
Convolutional Neural Networks in TensorFlow:
- From: Coursera
- Skills Learned: CNN
Introduction to TensorFlow for AI, ML, DL:
- From: Coursera
- Skills Learned: Computer Vision, TensorFlow, Machine Learning

Completed a 6-month internship (Apr 2023 - Oct 2023) at Samsung Prism and worked on the project “Corrupted Image Detection” using Deep Learning
Completed a 6-month internship at Samsung in the domain of Generative AI (Oct 2023 – Apr 2024)
Completed a 1-month internship at Thoughtware Analytics from (Jan 2024 – Feb 2024)
Completed a 3-month internship at Vaisesika from (Apr 2024 – Jun 2024)
Currently working as Junior AI Engineer at Vaisesika from (Jun 2024)

Corrupted Image Detection (Samsung):
- Classify images based on whether they are corrupted or not
- PyTorch framework was used with various CNN models such as TinyVGG, AlexNet, ResNetV1, and MobileNetV1 for mobile applications
- DistributedDataParallel (DDP) was utilized to distribute the training workload across multiple GPUs
Surgery Mortality Prediction:
- Determine whether patients can survive a particular surgery based on vitals
- Used PyTorch with algorithms like Random Forest, CNN, and DNN
Quality Assurance Framework for AI-generated Images (Samsung):
- Gathered a collection of image quality assurance metrics for assessing AI-generated images
Language to Language Models on the Opus Books Dataset:
- Constructed a transformer and trained it on a translation task
- Used PyTorch transformers with models like the standard transformer model, t5-small, nllb, and mbart
- Fully Sharded Data Parallel (FSDP) was utilized to train the model on multiple GPUs