Deep Learning

Period: 2021 – 2022
Scope: Deep Learning (Vision & Audio)
Tools: TensorFlow · PyTorch

Overview

Designed and deployed deep learning systems across medical imaging, microscopy analysis, and speech processing, with a strong focus on training pipelines, model optimization, and inference systems.

This role emphasized:

• end-to-end model development (data → training → deployment)
• handling real-world noisy datasets
• balancing accuracy, performance, and scalability

Key Contributions

• Built and trained segmentation models (semantic & instance) for medical and microscopy data
• Developed audio models for speaker recognition and voice cloning
• Designed data pipelines and preprocessing workflows
• Optimized models for inference efficiency and deployment
• Worked across TensorFlow and PyTorch ecosystems

Works

• Semantic Segmentation (Medical Imaging)
• Instance Segmentation (Microscopy)
• Speaker Recognition
• Voice Clone TTS
• WRB Crack Detection

Tech Focus

TensorFlow · PyTorch · UNet · Mask R-CNN · CNN · Audio Processing · Data Pipelines

Impact

• Delivered high-accuracy models for medical and scientific imaging
• Built reusable pipelines for training and inference
• Expanded capability into multi-modal AI (vision + audio)
• Strengthened expertise in production-ready deep learning systems