AMIT SHELKE

• Project 1
• Description: Accounting software that pre processes invoices
• Technologies: Python, MongoDB, NLP, Spacy NER, FastAPI
• 1. Collaborated with senior software start to produce new software solutions.
• 2. Contributed ideas and suggestions in team meetings and delivered updates on deadlines, designs, and enhancements.
• 3. Produced advanced and clean code for diverse application.
• 4. Implemented API service to communicate with NER model and Integrated with the application
• 5. Implemented Mongo DB aggregation pipelines

• Project 2
• Description: OMR Extraction API using ML
• Technologies: Python, MongoDB, FastAPI, OPENCV
• 1. Created an FastAPI using python to extract OMR data.
• 2. Implemented OMR extraction model using OpenCV2 library using object detection techniques.

• Project 3
• Description: Model development and Accuracy Improvement
• Technologies: Python, Pandas, Numpy, Scikit-learn, NLTK, Spacy, NLP, CNN, transformers, TenserFlow
• 1. Creating Data sets using Domain-specific data stored in Databases.
• 2. Implementing modules to clean data set.
• 3. Creating NER models using Domain-specific data using Spacy CNN & TransformerNER models.
• 4. Creating pipe lines to create models automatically.
• 5. Exploring AI models and AIML concepts

• Project 4
• Description: Face Recognition Model
• Technologies: Python, Pandas, Numpy, OpenCV, DL, TenserFlow, Docker, mongodb, FastAPI
• 1. Collaborated with seniors to create an architecture for the application
• 2. Implemented POC's for the application.
• 3. Built a Face Recognition API using Python, FastAPI, OpenCV, and face_recognition for facial detection and verification.
• 4. Designed REST endpoints to detect, encode, and match faces from image input.
• 5. Integrated NumPy for facial embeddings and stored data in MongoDB for persistent face records.
• 6. Containerized the application using Docker for easy deployment and scalability.
• 7. Documented APIs with Swagger UI and deployed locally and on cloud-ready environments.

• Project 5
• Description: Emma Healthcare chatbot
• Technologies: Python, openAI API, Docker, mongodb, FastAPI
• 1. Collaborated with senior software start to produce new software solutions.
• 2. Implemented Database schemas.
• 3. Built RESTful APIs for doctors, patients, appointments, analytics, and healthcare calls.
• 4. Integrated OpenAI for smart healthcare-related features.

• Project 6
• Description: Interview chatbot
• Technologies: Python, openAI API, Docker, mongodb, FastAPI, Langchain, LLM model
• 1. Built an AI-based Interview Bot using LangChain, OpenAI, and Streamlit.
• 2. Created a chatbot to simulate interview questions and answers.
• 3. Used LLMChain and PromptTemplate to manage conversation flow.
• 4. Built a user-friendly web interface using Streamlit.
• 4. Dockerized the project for easy deployment and environment consistency

• Project: Health Insurance Cross Sell Prediction
• Domain: Insurance
• 1. Identifying the customers whether they are interested in Vehicle Insurance provided by the company. The aim is to increase the revenue of the company
• 2. Actively involved in daily standup calls.
• 3. Analyzed & Filtered rows of data sent by client in multiple batches.
• 4. Optimized & Pre-processed data before building a classification model.
• 5. Performed Feature engineering on data using Python libraries like Numpy, Pandas, Seaborn.
• 6. Build and Trained the model using Random forest Classification and XGB classification.
• 7. Analyzed model prediction accuracy using Classification Report, confusion Matrix, AUC Score.

UI Trainee