AI Research Intern @ Metapix.me

Mar 2023 - Oct 2023

  1. Developed and implemented cutting-edge text-to-image and image-to-image stable diffusion models, resulting in a 30% improvement in user experience and engagement.

  2. Successfully optimized the stability and performance of the diffusion models through iterative testing and refining,achieving an impressive 80% accuracy rate in image conversion tasks.

  3. Spearheaded the creation and implementation of innovative image processing algorithms, resulting in a 30% increase in user satisfaction and participation for text-to-image and picture-to-image apps.

Machine Learning Intern @ Suvidha Foundation

Jul 2023 - Aug 2023

  • Fine tuned BERT model for text summarization, resulting in a 20% improvement in summarization accuracy compared to previous models.

  • Integrated the model into the company's existing text summarizing pipeline, achieving a 30% reduction in processing time and enhancing overall efficiency.

  • Collaborated with cross-functional teams to collect input on the adopted BERT model

Projects

LLM chatbot

Fine-tuned Llama-2 using PEFT and LoRA on a Question-Answer dataset to act as a customer support agent.

GitHub Learn More

Image Segmentation

Developed a U-Net convolutional neural network using the Pix2Pix model for up-sampling and the MobileNetV2 for down-sampling, achieving 93.77% training accuracy and 88.97% validation accuracy.

GitHub Learn More

Style Transfer

Implemented Neural Style Transfer using the VGG16 and VGG19 models. Performed a comraritive study of the performance of VGG16, VGG19 and ResNet-50 models in Neural Style Transfer.

GitHub huggingface Learn More