Conducted research on Multi-Agent Reinforcement Learning, employing Convolutional Neural Networks and Deep Deterministic Policy Gradient using CARLA to evaluate and understand agent behavior in a simulated environmentDeveloped software for a model car with Jetson TX2 and Arduino, creating data packets from RGB and event cameras, LiDAR, and IMU, and optimizing models using TensorRT with configurable Docker images for rapid deploymentImplemented Gaussian splatting and deployed it on the Nautilus cluster, utilizing fast gradient descent to create realtime 3D representations of a vehicle’s surroundings from SfM video clips, improving safety for autonomous vehicles

Developed an internal tool for real-time foreign key updates and LLM-driven descriptions for 2000+ tracking metricsAutomated CSV/JSON-to-DBML documentation in the customer frontend with Pandas and dbml-sync.Implemented B-tree indexing in PostgreSQL to optimize data retrieval and search operations in large databasesIntegrated CI/CD pipelines using Jenkins and Docker to automate testing and deployment of internal data tools

Built an online LMS using Flask and MySQL, integrating LangChain for feedback and serving 1,000+ monthly usersManaged operations for 9 branches, implementing Agile practices and overseeing technology support for 1,200 studentsDirected a 500-participant hackathon, securing sponsorships and guest speakers, awarding $2,000 in prizes

Presidential Volunteer Service Award for completing over 200 hours of service hours during the year 2020.Created Irvington CTI (Coding tomorrow) website (https://irvingtoncti.org) for my branch from scratch - Uses HTML, PHP, SASS, JS, Bootstrap. Added google analytics so that our publicity team could get detailed statistics on viewership after pub sessionsCreated and maintained the Irvington CTI discord server(https://discord.gg/AhHxqPeb) with an active community of over 180 membersHelped organize and judge the first-ever Irvington CTI hackathon with over 200 participants and $2000 in prizes.Held sessions for new volunteers and taught them essential teaching skillsCreated curriculum for new courses like Algorithmic Programming with Python and Intro to Arduino.Updated existing courses to be more engaging with our growing communityTaught over 200 students in courses like Beginner python and Advanced python

Worked as a mentor teaching over 100 students over a period of 6 months. Taught both beginner course and Advanced python course. Received many positive reviews from satisfied parents of students who attended my classes.Worked as an admin volunteer - managing grades and feedback for homework and tests

Developed machine learning and NLP applications to enhance personalized learning for individuals with learning disabilities, earning recognition from the California State AssemblyDesigned and built interactive product web demos for LLM prototypes and analyzed user journeys and clickstream data using A/B testing and data-driven optimizations to improve UX and conversion rates

2 terms at ASDRP - worked under Dr. Phil Mui (SVP Technology, Salesforce) Wrote a research paper on Detecting Police Bias in the American police force. Created and trained a machine learning model using data from American police killings to detect racial bias within the police force. Worked on data visualization, curation, and validation. Accepted into IEEE MIT Undergraduate Research Technology Conference, indexed into IEEE Xplore, and published to ASDRP communications third edition: https://drive.google.com/file/d/1pCGiu0VvarR10frh9nhUyfMK6Dp8bs2d/view?usp=sharing Learned libraries like pandas, scikit-learn, and matplotlib Worked pandas to clean datasets and validate data returned by our ML model Used matplotlib to visualize the cleaned data for use in the research paperUsed Scikit Learn to create and train a decision tree machine learning model to conduct our research onThe technology used - Python, Scikit learn, Matplotlib, numpy, pandas, pyscopg2, used google colab as our IDE, imblearn