Amit Jaiswal

docedgeAgent:-

• Created a robust facematch model from scratch using PyTorch and deep learning, showcasing expertise in computer vision and model development.
• Devised and implemented a creative solution for extracting data from invoices using NLP and NER techniques, incorporating LayoutLM for heightened accuracy.
• Collaborated across teams to deploy machine learning models in a production environment, ensuring seamless scalability and operational efficiency.

docedgeStudio:-

• Guided the development of the docedgeStudio model training tool, streamlining model creation and fostering better collaboration within the team.

EAHOMECARE

• Built a user-friendly chatbot using RASA 2.0 for hassle-free appointment bookings.
• Created a smart Q&A system with BERT to quickly find and provide information.
• Developed a COVID symptom severity predictor using Random Forest, aiding in timely patient care.
• Used K-means for a targeted customer segmentation model, optimizing engagement for vehicle insurance sales.
• Established a data pipeline in Python to efficiently gather information from various sources, including databases.
• Led a Proof of Concept for a recommendation engine, suggesting restaurants and gift items for concierge services.

PCB Component defect detection and failure analysis:-

• Perform data annotation by creating a bounding box for each object in an image using Labelimg.
• Perform data augmentation to increase the data size.
• Trained the yolov3 using pytorch on image data.
• Created an object detection model which detect missinghole, spurs, open circuit etc.
• Expose the model as web service using flask.

CATSCAN Web Platfrom:-

• Created a image classification model to predict relevant and irrelevant images using Resnet50.
• Created the data extraction pipeline from images using python, image processing using openCV and OCR.
• Perform data preprocessing using pandas and convert data into structured format and dumping in database.
• Created a sentiment detection model which predict user behavior on product review data using LSTM.
• Perform EDA to generate usefull insights.