Harshvardhan Rijhwani

Developed a travel itinerary generator that personalizes trips based on user location and duration, integrating interactive maps for visual
clarity.
Implemented itinerary management features, allowing seamless access to and modification of previous travel plans.
Enhanced user experience with an organized presentation of itineraries and integrated a chatbot for immediate user support.
Developed MERN webapp utilizing OpenAI API for itinerary generation and chatbot functionality.

Leukemia cell classification using image processing of more than 3000+ microscopic cell images of 4 categories
Implemented CNN models which outputs a blood cell image as cancerous/non-cancerous & also classifies the cell into different types
of leukemia Model based on mobilenetv2 CNN architecture and outputs an accuracy of 97%
Deployed model via Fast API as a robust backend server and a Frontend in HTML and CSS.

Developed an AI-powered Streamlit app to evaluate resumes against job descriptions using a configurable local LLM (via Ollama).
Implemented features for resume summarization, skill gap analysis, and job-fit percentage scoring with structured GPT-style prompts.
Built a section-wise heatmap using Plotly to visualize resume relevance (Skills, Experience, Projects, Education).
Integrated PDF parsing, JSON parsing, and real-time response streaming with modular prompt handling and model selection.

Organized & volunteered in flagship event of committee, TSEC Hackathon 2023

Conducted comparative analysis of ARIMA, LSTM, and SVM algorithms for stock price prediction across four stocks and three
timeframes: all time, COVID (2020-2021), and five years.; Identified ARIMA as the most effective model for stock price prediction
based on empirical comparisons and statistical metrics.