Daksh Khandelwal

Agentic AI esports coach that analyzes gameplay video with Gemini Vision and a LangGraph multi-agent pipeline. Extracts frames, builds a chronological timeline, and evaluates positioning, mechanics, and decision-making — producing a personalized coaching report with strengths, mistakes, and drills via a Streamlit dashboard.

ResearchMind is a multi-agent AI research system built with LangChain and Mistral AI (mistral-medium-3.5). Four agents — Search, Reader, Writer, and Critic — collaborate to search the web, extract content, draft a structured report, and critique it. Includes a Streamlit UI and CLI, powered by Tavily search and BeautifulSoup scraping.

🏎️ F1 Race Analytics is a Streamlit-based machine learning application that analyzes Formula 1 race data and predicts overtaking opportunities. It features interactive dashboards, race statistics, probability analysis, leader tracking, and visual insights to help users explore race performance and strategy decisions.

Built an NLP-based sentiment analysis model on Amazon Fine Food Reviews using TF-IDF and machine learning algorithms. Applied preprocessing techniques like tokenization and stopword removal to improve feature representation and achieved over 85% prediction accuracy on review sentiments.

A Python-based NLP autocorrect system that detects and corrects misspelled words using edit-distance techniques and word probability prediction. Includes a Streamlit interface for real-time text correction.

This game is a 2D wave-based survival game built in Godot Engine. You play as a lone hero battling endless waves of bats and skeletons, surviving as long as possible while racking up a high score.

This is a to do list app which tracks your progress, lets you add delete and edit tasks.

a weather app that uses openWeatherMap API to get real time weather