With some of my free time between graduation and beginning my full-time job, I decided to learn more about how pairwise comparisons can be used to rank items in a pool without performing all possible pairwise comparisons. Of course, one of the most famous and well-known examples of this is the Chess ELO algorithm. Feel free to click the link above to see a simple animated visualization of this algorithm in real time. Notably, while this can certainly be used for creating good matchups in games, the uses of this algorithm go far beyond games. That's why I also created a Python library to facilitate more modern/advanced versions of this algorithm in a reusable way with generics. This library can be found on my GitHub.

During my final semester at UTD, I was offered a full-time position at Walmart Global Tech as a Software Engineer. I was thrilled to continue my journey with WGT as I had a blast at the internship. I will begin this offer in February 2025.

After three and a half years of learning and fine tuning my skills in Computer Science at UTD, I finally graduated with a 4.0 GPA (Suma Cum Laude) in Computer Science. While leaving UTD is bittersweet, I am so excited for my next step in the professional world.

During my final semester at the University of Texas at Dallas, I had the brilliant opportunity to work with Spark Biomedical, a large startup blossoming with ~$20m in funding. Specifically, Spark Biomedical has been leading the world in creating the first non-invasive neurostimulation devices potentially used for solving numerous health issues that millions of people face. Personally, my team and I were tasked with creating a whole new platform to help facilitate fully online clinical trials in a safe and secure manner. Specifically, I worked on creating the central API, database schema, authentication scheme, and supported my teammates in creating platform for clinicians as well as a client app to interface with the devices. Overall, after presenting our solution at the end of the semester in front of industry leaders, we won first place in our graduating class for our project.

Over the past summer, I spent my time in Bentonville, Arkansas working at Walmart Global Tech alongside some of the brightest minds I have ever met. While many people may not expect Walmart to have such a large tech workforce, I quickly learned that Walmart had some of the most unique and fun problems to solve in the entire industry. Notably, Walmart had a whole new level of scale form anything I had ever seen. I had a blast learning all about tools, both open source and proprietary, that were able to handle this scale and exceed performance expectations. Through the use of Springboot, Kubernetes, and more, I even created my own project that saved hours every week for the business teams and presenetd this project to the Vice President over our organization.

After spending hours of our free time tailoring our resumes to fit the relevant experiences on our resumes for each job we were applying to, myself and Timothy Naumov decided to create a tool to automate the entire process. Hence, Chisl was born. Originally starting as our independent study at UTD, Chisl was created using custom AI models trained to understand and express your best skills and experiences for every single job in seconds, allowing you to click one button and have a resume downloaded to your computer in less than 30 seconds with your exact experience front and center for the job you're applying to. Today, Chisl is in a Closed Beta through a collaboration with the UTD career center to introduce 100s of students to our powerful new tool. While we are still working diligently to improve Chisl, we already have tons of users that already love and use our tool.

During my senior year of college, I was one of only five students selected to be a consultant at 180 Degrees Consulting out of over 100 applicants. I was placed on a team of four other consultants and we were tasked with helping a local nonprofit, Johnathan's Place, grow through a reimagining of their tech stack. We were able to provide them with valuable tools and insights that helped automate many of the processes that previously took up a tremendous amount of time. This was a wonderful experience that I will never forget, and I am extremely grateful for the opportunity to have been a part of it. Having the chance to help so many people in need was truly a life-changing experience.

One of HackPrinceton 2023's highest ranking projects, Redwood is a truly novel project that made it to the final round of judging at HackPrinceton and is aimed at changing the way we all perceive information in documents such as textbooks or research papers. Using FastAPI, Svelte, Spacy, and OpenAI, Redwood tasks in a PDF of any size and creates a "Learning Tree" of it. Our "Learning Trees" are a graph of nodes that contain important topics throughout the PDF. By looking at the directed connections between these topics, users are able to quickly understand the foundational information for higher-level concepts in their documents. For example, if we were to make a learning tree of a Database Textbook, some things like indices may be crucial to understanding certain SQL statements, so there would be a connection between these nodes. On top of this, if users double click on a node in the graph, they will get an immediate summary of the topic using Retrieval QA with GPT-3.5-Turbo. Because this isn't an interactive summary, however, we also provided an AI-agent that acts a tutor to allow users to better understand the given topic. The tutor is also given the ability to perform Retrieval QA as well, so that it's always up-to-date and has correct information from within the document you are viewing itself. Finally, this novel approach to learning creates a number of powerful use cases, such as topological sorts (which are possible because the learning trees are DAGs) that would allow a user to quickly understand the exact information they would need to learn to fully understand a later concept and this would even be able to tell them the exact order in which they need to learn the information as well. Overall, this is one of my favorite projects I've ever created primarily due to the novelty of the idea.

Created as a project submission to HackUTD X, Start Smart is one of my most fun and ambitious projects that I've created in a while. Start Smart is specifically designed as a way to help small businesses grow in size without having to risk large amounts of money on industry experts to solve issues, especially since many small businesses aren't able to afford employees for market research, HR purposes, and much more. To solve this issue, we created Start Smart: a game environment built in vanilla JavaScript and React that enables users to talk to many varied specialized AI agents that each have their own abilities. For example, the Research agent is able to search through company documents and other PDFs that the user uploads (such as research papers) and the agent is able to query these documents with embeddings dynamically in conversation and use the context given by the documents to provide the user with insights and information. On the other hand, the Financial agent is able to scrape the web for articles relevant to the businesses finance and provide more specific insight on those articles to the user and how they apply to the user's business. Finally, our team created a Kanban board that allows users to specify different tasks that are relevant to their business, and the phases that they need to be completed in. On top of this, each AI agent is able to create tasks for this Kanban board as well as read from it to gain a greater insight on the overall project. These agents are even able to remember the tasks they've created, as well as track the tasks throughout the board itself, meaning that they will know immediately if you've completed a relevant task, and will be able to talk about it dynamically.

Created as a submission to HackTX, Mental Foundation is a full-stack application built to slowly break down the stigmas of mental health in the construction industry. Through the use of Streamlit, Flask/Python, Svelte, OpenAI, Tensorflow, SQLite, and Twilio, Mental Foundation would provide daily and weekly surveys for construction workers about their day-to-day life. Through these surveys, sentiment analysis was used to extract a mental "score" for each submission, and then these scores were processed by our predictive AI model on the server to determine if there was any chance that mental health resources would be needed. The forms were sent over text, and could be accessed on any device, while the managers of the workers would be provided a Streamlit dashboard which they could use to visualize the trends of their overall worker group. Using this dashboard, managers could quickly gain some insights on the reasons for mental issues. One great example of this is if we saw the entire group of workers having a degradation of mental health, the source could quickly be determined to be factors like overworking or environmental conditions, both of which are extremely common issues in the construction industry.

HackUTA’s second place overall winner which took advantage of Google Voice, multiple LLM agents, Virtual Audio Cables, State Farm’s Drive Safe data, Python, Twilio, and Elevenlabs to create a fully functional AI phone representative that was able to text users whenever their cars were in need of preventative maintenance, and then personally call users to have a conversation with them regarding any information about either the user’s car, the maintenance required, or any other questions that need answering for the user. On top of all of this, Amelia was also able to query and use the user's data through State Farm's Drive Safe app to only provide the most relevant and personalized assistance to users. Uniquely, because Amelia was entirely an over-the-phone assistant, you could even utilize her assistance on a flip phone or pay phone.

HackRICE’s second place winner for SLB’s climate change challenge which utilized React Native, Sveltekit, and Flask to create a fullstack mobile app and website that helped users optimize their daily carbon footprint through gamification and AI. All the users would have to do to get started is speak their daily schedule into their phone, at which point their voice would be transcribed and converted into a climate friendly schedule. Some other features included an incentive system for users in the form of points for reducing their carbon footprint, leaderboards for user groups, and so much more.

Model Maintenance was a project created for HackSMU V which was focused on building a interface for realestate agents and landlords to help them better manage their properties using a 3D interface through a webpage that also included machine learning to accurately predict when various different systems would be in need of maintenance. Utilizing Sveltekit, Threlte, Flask, and TensorFlow, we were able to satisfy the needs of any realestate manager looking for a smarter and more efficient way to manage their properties.

As someone who generally becomes quite passionate about my papers, I decided to create a tool which helps me make sure that all of the work that I do is my own. To do this, I created QuoteSearch, a tool that takes in the research papers you are reading while you write your paper, and then allows you to search for all the most relevant paragraphs containing concepts you are using in your paper. In doing so, you are able to quickly and easily cite the specific parts of the research paper you need while avoiding re-reading through hundreds of pages to find specific quotes.

As someone who generally uses WSL, I get quite confused with the various different commands for different OSes. Because of this, I made a simple cross-platform commandline tool that allows you to type in a description of what you're doing and then it will translate it into the correct command for your Operating System.

This is a simple tool that I created to allow people to quickly and easily create functions to be used in the OpenAI Chat Completions API. The front-end uses Sveltekit and the entire webapp is static so that it can be freely hosted on GitHub Pages to live on forever.

This was a simluation I created as a fun and educational way to visualize darwinian evolution. Each fish has a speed it is born with, and that speed is mutated whenever a fish reproduces. The more red a fish is, the faster it is. Click the title above to see for yourself, and have fun!

Since I started working on my portfolio website, I figured that I should also add some of my other projects to it. This was the third, a JavaScript version of my old Dijkstra's Visualizer. It's written in vanilla JavaScript, and uses Svelte for the frontend.

Since the BubbleSort Visualizer was so slow, I decided to improve on my next one and do BucketSort, which has the best time complexity of any sorting algorithm for this small of integer inputs. Since there are only 100 possible values, the time complexity is O(100). This was written in vanilla JavaScript, and uses Svelte for the frontend.

To start off my JavaScript portfolio, I decided to create a simple BubbleSort visualizer. This was written in vanilla JavaScript, and uses Svelte for the frontend. I plan on adding more algorithms to this in the future, but for now this is the start.

With some free time on my hands between my exams and internship, I decided to work on bettering my professional presentation. Thus, this resume was born.

During HackAI 2023, two other teammates and I created a fullstack suite of tools to revolutionize the stadium experience. This came in the form of a "betting" system that allowed users to gain points throughout the match and then stake their points on AI generated bets that were based on the current state of the game, as read from the announcers. To further enhance these bets, we also created a machine learning model that was able to predict (with ~75% accuracy) the outcome of a basketball match. This allowed us to make more realistic bets and even tweak the odds of the bets to further captivate the users. We used React Native for our client app, Svelte for our admin app, and Flask for our backend.

During my fourth Hackathon, Axxess' first inaugural Hackathon, I created an app called MedMate with three teammates. This app was designed to be an all-in-one medical app that allowed users to track their medical history, get reminders for their medications, and even get personalized descriptions of drug side effects. On top of all of this, we added a Smart Transcription service that allowed users to record their doctors appointments (with the doctor's consent) and get a translated transcription (supported around 40 languages) of the appointment in real time. The Smart Transcription would then analyze the appointment and extract only the most important information and highlights of the transcript. We created this app to be cross-platform using React Native, and used Flask for our backend. We won first place overall in the Hackathon.

During my third Hackathon, WeHack 2023, I created a webapp called Site In-Sight with two other people. This was a webapp that allowed users to search for an area, get listings, and determine the most profitable realestate for sale using a machine learning model. We used Flask for out backend, Svelte for our frontend, and Google Cloud for our data. We won first place in the CBRE Machine Learning challenge.

After just one year of being a member of the Google Developer Student Club, I was so excited to be pulled aside by the president at the time and told that I was going to receive his endorsement for being the next GDSC president. After going through the Google interview process, I received the role of president and led the club to have some of the biggest collaborations the club had seen. This included an interview with the co-founder and former CEO of GitHub Chris Wanstrath, a recruiting event with Epsilon, and much more.

After my Brains4Drones internship ended, I ended up having a lot of free time, so I began looking for internships once more. Fortunately, I had the pleasure of getting an internship with BenefitMall, a subsidiary of Truist. I worked on a team of 5 other people, and we were responsible for supporting the Production webapp EmployerFocus. Early on in my internship, I was able to make large contributions to the codebase, and completely rehauled the W-2 Report system, increasing the speed up to 32x. During this internship, the skills I developed consisted of C#, SQL, and Azure.

During my second ever Hackathon, HackUTD 2022, I created an app using Flutter that takes in a user's financial transactions and uses their data to better understand how spending money makes them feel, which allows for diagnosing habits and improving them. I worked with a team of two other people.

After attending a couple of meeting with our local Google Developer Student Club, I realized that GDSC was the perfect place for me to continue my love of teaching. After going through the interview process, I was ecstatic to learn that I had been chosen to work on the Technical Committee creating and running technical workshops for my fellow colleagues. It was here that I ran my first solo workshop (on data visualization and statistics with Python).

During my first ever Hackathon, HackUTD 2021, I worked with two other teammates to create a python desktop application that allows students to quickly and easily compare professors at UTD. This included scraping data from RateMyProfessor combined with UTD grades. We then used a custom algorithm to determine the best professor for a student based on their preferences.

After only a short while of attending classes at UTD, I was able to secure an internship with Brains4Drones, a high-growth startup that was focused on providing AI tools and intelligent drones to utility companies. While working here, I learned a lot about Linux, OpenVINO, Tensorflow, YoloV5, and more. Since I was also stayed on for a year, I also had seniority and the pleasure of leading a team on the creation of the embedded file system architecture for the drone. I also committed to Ultralytics' YoloV5 repository and helped iron out bugs in their code. During this internship I refined my Linux, Python, and C++ skills.

With a few smaller wins around my belt at the highschool level, I was ready to enter the new "big pond" and begin my Computer Science degree at UTD. During my first class, I had my love for programming reaffirmed once again as I learned C++ formally for the first time and entered a whole new world of memory management.

With the extra time at my computer due to the COVID-19 pandemic, I decided to take the time and try something new. I got to work on a graphics engine, written entirely in Java with no libraries. This meant that everything from the matrix multiplication to the rendering was done by me. I learned a lot about linear algebra and 3D graphics, and I am very proud of the result.

On top of placing in district previously, we were also invited to an invitational, and placed first there as well.

Unfortunately, with the COVID-19 pandemic, we were unable to compete in the regionals and statewide competitions. However, we were able to compete online for district, and placed first place. This time, since it was my senior year, I had the honor of not only becoming captain of our Computer Science UIL team, but I also had the opportunity to teach our team everything that I had learned from the ground up, as we went out of our way to recruit students that could carry the torch in the future. I am proud to say that many of the students I taught have now gone on to both win competitions and even secure impressive internships at large tech companies.

I was extremely fortunate to have a school that supported me and my small team of programmers. We were able to compete in the Texas Computer Science UIL Competition. We ended up placing first in our district, and then in the top ten for our region. This was a huge accomplishment for me and my team.

After having practiced Computer Science for a long time as a hobby (and growing my passion) I decided that I wanted to spread my knowledge. This came in the form of a job at Code Ninjas, where I mentored many different students, wrote curriculum, and taught classes.

I took my first formal programming class in high school. I learned the basics of Java and C++. I also learned to put a name to many of the concepts that I had been learning-through-doing over the past 8 years. I was hooked.

I started programming in Java when I was 10 years old. I was inspired by the game Minecraft and wanted to make my own mods for it. I started with a simple mod that added a new block to the game. This quickly progressed to a new biome, and then new weapons and ore.