About Me

Hey there, I'm Mohit Soni! Welcome to my little corner of the internet! If you're here, you're probably wondering who I am and what I do. Well, buckle up, because I'm about to take you on a whirlwind tour of my journey so far! I'm currently a Computer Science Major at Cornell University. My academic journey is filled with algorithms, data structures, and a whole lot of late-night coding sessions.

Recently, during the summer of 2024, I worked as an AI Software Developer Intern at BrickRed Systems. I've been part of an amazing team of 25 people, working on EchoMD, an AI-powered platform that's shaking up the healthcare industry. My job? Making sure that doctors can spend more time with patients and less time on paperwork. I developed and fine-tuned some nifty NLP algorithms using Python libraries like NLTK and SpaCy, all to transcribe doctor-patient dialogues in real-time. TensorFlow was my trusty sidekick in training and deploying these models. I also dabbled in backend magic with Django and Flask, whipping up RESTful APIs to keep everything running smoothly. The result? A 30% reduction in documentation time and a 25% increase in record accuracy.

Not only that, under the mentorship of Professor Madhur Srivastava at the Signal Science Lab at Cornell University, I’m leading a research initiative to develop an algorithm that analyzes cricket batsmen's techniques based on video data. Think of it as a tech-driven coach that scrutinizes every stance, footwork, and bat swing. I crafted an innovative algorithm for video segmentation using pose estimation and dynamic time warping, ensuring we capture only the most relevant frames.

But when I'm not busy working on my research project or internship, I like to let loose and have a little fun. That's where Kaggle competitions come in - it's the perfect way for me to show off my tech prowess and flex my analytical muscles. Plus, I'm always working on my own personal projects - because why not?

Though I may look like your typical tech nerd, but I promise I'm not just a coding machine. I'm also a serious cricket player! I'm the President of the Cornell Cricket Club and I have been a member of the U19 team at the National level. Being part of this team sport has taught me so much about leadership, teamwork, and perseverance. From serving as the Captain of the Cornell Cricket Team playing for regional teams around Seattle, I've learned to motivate and communicate effectively with my teammates, identify their strengths and weaknesses, and help them improve their game.

Cricket isn't my only passion – I've recently become an avid chess enthusiast too! I got hooked on the game during my senior year of high school, and now I play consistently, fitting in 2-3 games every day. I love the thrill of blitz and the strategic depth of rapid chess. My dedication has paid off with a rapid rating of 2000 and a blitz rating of 1700. Chess has taught me patience, focus, and strategic thinking, making it a perfect complement to my interests and a great way to stay sharp and competitive.

My Projects

Cricket Weakness Detector Model

I did this project with UW Professor Sundar Srinivasan, where I used machine learning to predict the chances of a Cricket batsman getting out based on the field, ball trajectory, and the batsman's historical statistics. I worked on this project further in 12th grade to improve my results.

ML Model Comparison App

By using the Streamlit library, I built a web application that allows users to train several Machine Learning models on different datasets and quickly assess the efficacy of the models side-by-side WITHOUT any coding or Machine Learning experience.

The Complete 2022 Web Development Bootcamp

By taking this Udemy course in August 2022, I became familiar with frontend frameworks, like Bootstrap and React, and backend technologies, like Node.js and MongoDB. I also completed 18 projects that incrementally increased in complexity which allowed me to solidify my skills.

Complete ML & DS Bootcamp

By taking this Udemy course in December 2022, I became familiar with ML concepts from data collection to hyperparameter tuning. I also completed 3 projects (covering regression, classification, and computer vision) giving me the skills to pursue more in-depth projects.

Kaggle Microbusiness Density Forecasting

In this competition, I used my data analysis and ML skills to predict the density of microbusinesses in different regions based on historical data. I was exposed to the difficulties of dealing with time-series data and got experience with prediction models such as ARIMA and SARIMA.

Chess Game Accuracy Predictor

I'm currently working on a fascinating project where I explore the potential of neural networks to predict the accuracy of a chess game, all without relying on a traditional chess engine. The idea is to harness the power of machine learning to analyze game data and provide accuracy insights, opening up new possibilities for chess analysis.