This repository contains a collection of AI and NLP projects developed during my professional internship. Each project demonstrates different aspects of AI application development, from text summarization to multilingual chatbots.
An intelligent text summarization tool powered by BERT that automatically extracts the most important sentences from documents to create concise summaries.
Key Technologies: Flask, React, PyTorch, Hugging Face Transformers, NetworkX, PyPDF2, pdfplumber, TailwindCSS, Vite
A comprehensive system for chatbot interactions with an analytics dashboard to track user engagement, satisfaction ratings, and conversation topics.
Key Technologies: Python, Ollama (Cogito:8b), Plotly, Dash, Pandas, SQLite, Chart.js, Bootstrap
A tool for generating articles using different open-source LLMs (Mistral, Qwen2.5, Llama3.1) and comparing their performance.
Key Technologies: Python, Ollama, Flask, NLTK, spaCy, Matplotlib, scikit-learn, TextBlob, Jinja2
A chatbot that can understand both text and image inputs, providing seamless multi-modal interactions.
Key Technologies: FastAPI, React, Ollama (llama3.2-vision:11b, qwen2.5:7b), Pillow, Material-UI, Axios
A medical question-answering chatbot that provides appropriate and relevant medical information using retrieval-based methods and LLMs.
Key Technologies: FastAPI, React, Ollama (Meditron 7B, Llama 3.1 8B), MedQuAD Dataset, FAISS, Pandas, NumPy, Tailwind CSS, TypeScript
An AI-powered system that monitors YouTube channels, extracts transcripts, identifies key insights, and allows users to query this information through a chatbot interface.
Key Technologies: FastAPI, React, MongoDB, Pinecone, YouTube Data API, YouTube Transcript API, OpenAI GPT-4, LangChain, Pydantic
A sophisticated chatbot with sentiment analysis capabilities to recognize and respond appropriately to customer emotions.
Key Technologies: FastAPI, React, Hugging Face Transformers (RoBERTa), Ollama (Llama3.1:8b), Material-UI, Chart.js, Pydantic, WebSockets
A domain-specific expert chatbot using arXiv papers as a knowledge source, powered by fine-tuned Mistral 7B and retrieval-augmented generation.
Key Technologies: Python, Mistral 7B, Faiss, MongoDB, FastAPI, React, Docker, LangChain, Sentence-Transformers, PyPDF2, D3.js, React-Force-Graph
A multilingual chatbot supporting English, Hindi, Bengali, and Marathi with automatic language detection, translation, and transliteration capabilities.
Key Technologies: FastAPI, React, fastText, Ollama (LLaMA3.1-8B), IndicBART, IndicXlit, styled-components, Pydantic, Uvicorn
The projects in this repository utilize a variety of technologies:
- Backend Frameworks: FastAPI, Flask, Uvicorn
- Frontend: React, TypeScript, Vite, Material-UI, TailwindCSS, styled-components
- AI/ML: PyTorch, Transformers, Hugging Face models, LangChain, NLTK, spaCy
- LLMs: Mistral, Llama, Qwen, Meditron (via Ollama), OpenAI GPT-4
- Databases: MongoDB, Faiss, Pinecone, SQLite
- Vector Embeddings: Sentence-Transformers, all-mpnet-base-v2
- Visualization: Chart.js, D3.js, Plotly, Matplotlib
- Deployment: Docker, Docker Compose
- APIs: YouTube Data API, YouTube Transcript API
Through these projects, I've demonstrated proficiency in:
- Large Language Model Integration: Implementing and fine-tuning various LLMs for specific tasks
- Retrieval-Augmented Generation (RAG): Creating systems that combine knowledge bases with generative AI
- Multi-Modal AI: Building applications that process both text and image inputs
- Natural Language Processing: Implementing text summarization, sentiment analysis, and language detection
- Full-Stack Development: Creating end-to-end applications with modern frontend and backend technologies
- API Development: Designing and implementing RESTful APIs
- Database Design: Working with both traditional and vector databases
- Data Processing: Extracting, transforming, and analyzing structured and unstructured data
- UI/UX Design: Creating intuitive user interfaces for complex AI applications
- System Architecture: Designing scalable, modular systems with clear separation of concerns
Through this internship, I've gained valuable experience and insights:
- Practical AI Application: Moving beyond theoretical knowledge to build real-world AI applications
- Technical Integration: Combining multiple technologies and frameworks into cohesive systems
- Problem-Solving: Addressing challenges in AI implementation, from model selection to deployment
- Performance Optimization: Balancing model accuracy with computational efficiency
- User-Centered Design: Creating AI systems that are accessible and useful to end-users
- Ethical AI Development: Considering privacy, bias, and ethical implications in AI applications
- Project Management: Planning, executing, and documenting complex technical projects
- Continuous Learning: Adapting to rapidly evolving AI technologies and methodologies
Each project has its own README with specific setup instructions. Generally, the projects follow this pattern:
- Set up a Python environment
- Install dependencies from requirements.txt
- Install Ollama and pull required models (for LLM-based projects)
- Set up the backend server
- Set up the frontend application
This repository is for demonstration purposes. All projects are provided under the MIT License.