Himanshu Upadhyay

About

Generative AI Engineer with hands-on experience building and deploying LLM, RAG, and agentic AI applications using Python, LangChain, LangGraph, vector databases, and FastAPI. Built document Q&A, conversational AI, text-to-math, and multi-agent research systems with web research, structured outputs, and grounded responses.

Experience

• ML/AI Trainee
  Apna Time Tech Solutions
  Jan 2026 – Jul 2026

Education

• B.Tech
  Malout Institute of Management & Information Technology
  Computer Science & Engineering · Jul 2022 – Jul 2026

Skills

• Deep Learning
• FastAPI
• Python
• Machine Learning
• Prompt Engineering
• Streamlit
• TensorFlow
• LangChain
• Natural Language Processing
• RAG
• LLMs
• LangGraph
• Faiss
• Agentic AI Systems
• Generative AI (ChatGPT, Copilot, Gemini)
• Vector embeddings

Projects

• Next Word Prediction ↗
  TensorFlow, Deep Learning, NLP, Streamlit

  Built an LSTM-based next-word prediction model for sequence text generation and deployed it through an interactive Streamlit application.

• Twitter Sentiment Analysis ↗
  TensorFlow, Deep Learning, NLP, Streamlit

  Built an RNN-based sentiment classifier for tweet text and deployed it as a Streamlit application supporting user text input and inference.

• Movie Recommender System ↗
  Python, Streamlit

  Built a content-based recommender using feature-based similarity across movie metadata and deployed an interactive recommendation interface.

• GenAI Chatbot ↗
  Python, LangChain, Groq API, Prompt Engineering, Streamlit

  Developed a conversational LLM application using LangChain and Groq API, with prompt engineering for context-aware responses. Deployed an interactive Streamlit chat interface for user-driven conversations.

• RAG Chatbot with PDF & Chat History ↗
  Python, LangChain, RAG, Vector Embeddings, FAISS/ChromaDB, Streamlit

  Implemented a PDF RAG workflow covering document ingestion, chunking, embeddings, similarity retrieval, and grounded LLM response generation. Added persistent conversational history for multi-turn document Q&A and deployed the application through Streamlit.

• Text-to-Math Problem Solver ↗
  Python, LangChain, Groq API, Agent/Tool Calling, Streamlit

  Built an LLM-powered text-to-math application that interprets natural-language math problems and uses an agent/tool workflow to produce solutions. Integrated LangChain with the Groq API and deployed the application through Streamlit for interactive problem solving.

• Multi-Agent AI Research System ↗
  Python, LangChain, LangGraph, OpenAI, Tavily, BeautifulSoup, Trafilatura, Readability-lxml, Streamlit

  Built a multi-agent research pipeline with specialized Search, Reader, Writer, and Critic stages for automated web research and report generation. Integrated Tavily web search and multiple content-extraction fallbacks, then generated structured reports and evaluated output quality with a critic stage.

Courses & certifications

• ML/AI Traineeship Certification · Apna Time Tech Solutions