-Implemented a Stable Diffusion-based text-to-image generation pipeline from scratch using PyTorch, inspired
by the DDPM (Denoising Diffusion Probabilistic Models) paper.

• Designed and trained core architectural components such as U-Net, Variational Autoencoder (VAE), and CLIP
  for prompt-based image generation.
• Built an interactive Python interface to visualize input prompts and corresponding image outputs for real-time
  experimentation.
• Gained in-depth understanding of latent diffusion models (LDMs) and their application in modern generative
  AI workflows.

• Developed QA and summarization system from PDFs, web URLs, and YouTube videos.
  -Integrated OpenAI/Gemini LLMs for answer generation and LangChain for document processing.
• Enabled semantic search with FAISS and multilingual summarization (English/Hindi).

•Developed an AI-powered chatbot for Dreamadoration to automate user interactions and answer customer
queries in real time.
•Implemented NLP and LLM-based response generation to understand user questions and provide relevant,
contextual answers.
•Designed a retrieval-based pipeline using embeddings to fetch accurate information from company
documents and FAQs.
•Built the backend using Python and integrated it with a vector database to enable fast semantic search and
scalable responses.

•Built an AI-powered Creative Writing Analyzer that evaluates student writing for tone, structure, and
originality and provides real-time feedback.
•Implemented LLM-based analysis and embedding techniques to assess coherence, creativity, and writing
quality.
•Designed structured prompts to generate constructive feedback including strengths, improvement areas, and
suggestions.
•Built the backend using Python with AWS S3 for storage and MongoDB for managing writing data and
analysis results.

•Developed AI Chatbot solutions for businesses, enabling automated conversations for customer support, lead
generation, and user assistance on websites.
•Built chatbots that use natural language processing and LLM-based response generation to understand user
queries and provide contextual answers.
•Designed custom chatbot workflows for clients, allowing features like answering FAQs, capturing leads,
guiding users through services, and handling customer inquiries automatically.
•Integrated the chatbot with websites, databases, and business APIs to enable real-time responses and
improved customer engagement.

•Developed AI Voice Agent solutions for businesses, enabling automated phone-based conversations for
customer support, sales, and service inquiries.
•Built systems that use speech recognition, natural language understanding, and text-to-speech to allow users
to interact with AI agents through natural voice conversations.
•Designed custom voice workflows for clients, enabling features like answering customer queries, booking
appointments, lead qualification, and automated call handling.
•Integrated the voice agents with telephony systems and business APIs to enable real-time conversations and
automated business actions.