StudentDeveloperResearcher
Hello, I'm Rajit Sareen
CS @ University of Washington, developer and software engineer fueled by curiosity and grit to build.
My Story
I'm currently studying at the Paul G. Allen School of Computer Science at the University of Washington. My interests lie in computational mathematics and theoretical computer science, where I enjoy exploring the mathematical foundations of computing and algorithm analysis. I'm particularly drawn to complex problems that combine analytical thinking with practical implementation.
Milestones
My Journey
Software Lead, FIRST Robotics
Sep 2021 – Jun 2023
Led software development for an autonomous robot using Java and OpenCV, achieving a top-5 state ranking and developing a passion for computer science.
JavaOpenCVRoboticsLeadership
IT Intern, Richland School District
Jun 2023 – Aug 2023
Imaged and deployed over 500 computers, managed inventory, and provided technical support to staff, enhancing problem-solving and communication skills.
IT SupportSystem DeploymentTechnical Communication
University of Washington
Sep 2023 – Jun 2028
Pursuing a Bachelor of Science in Computer Science at the Paul G. Allen School, focusing on AI, systems, and computational mathematics. GPA: 3.75.
CSAI/MLSystemsAlgorithms
Participant, Husky Datathon
Apr 2024
Developed a machine learning model to predict company profit based on multi-product sales data, utilizing XGBoost for high accuracy.
Machine LearningXGBoostData Analysis
Undergraduate Research Developer, DAIS Lab
Jun 2024 – Jan 2026
Enhanced protein structure prediction through U-Net and diffusion-based generative AI models on cryo-EM data. Trained deep learning models with PyTorch on high-performance GPU clusters.
PyTorchDiffusion ModelsCryo-EMResearch
DubHacks Participant
Nov 2025
Competed at DubHacks and developed AppetizeAI, an AI-powered restaurant recommender and map interface using AWS Bedrock foundation models.
Artificial IntelligenceAWS BedrockHackathon
Deep Learning Researcher, ContextOpt
Jan 2026 – Present
Benchmarking 6 context optimization strategies (RAG, Summarization, LLMLingua, Hybrids) across LongBench, evaluating Llama-3.1-8B and Gemini-2.5-Pro on quality, token efficiency, and cost. Training an adaptive XGBoost and MLP selector to route queries to the optimal strategy.
LLM ResearchRAGXGBoostContext Optimization
My Skills
React
Next.js
TypeScript
Python
Node.js
Docker
AWS
JavaScript
FastAPI
MongoDB
SQL
PostgreSQL
Java
Git
PyTorch
Claude AI
AI / ML
Neural NetworksDiffusion ModelsRAGPrompt EngineeringMCP ServersAI AgentsLLM Fine-tuningComputer VisionXGBoost
My Certificates
Dell Tech Direct
Certified for Dell's Tech Direct program, gaining expertise in diagnosing and repairing Dell hardware.
SolidWorks
Certified in SolidWorks, demonstrating proficiency in 3D modeling and mechanical design.
AWS Cloud Practitioner
Certified as an AWS Cloud Practitioner, validating foundational knowledge of AWS cloud services.
My Projects
AppetizeAI
Django, PostgreSQL, AWS Bedrock
A full-stack restaurant discovery app supporting natural language search and real-time, context-aware recommendations using foundation models.
SmartFridge
React Native, FastAPI, PostgreSQL, Docker, YOLOv8
Mobile app identifying fridge items via YOLOv8 trained on 3K+ labeled images, with FastAPI backend, JWT auth, and Docker deployment.
Monte Carlo Monopoly
C++, Monte Carlo Simulation, Statistics
Monte Carlo simulation engine in C++ modeling 50K+ Monopoly games, analyzing expected value, variance, and bankruptcy probability.
Contact
Let's Talk
Interested in working together or have any questions?