Ayan Agarwal
Computer Engineering Student
About
- Major: Computer Engineering
- School: Georgia Institute of Technology
- Email: ayansagarwal@gmail.com
- Region: Atlanta, GA
- Citizenship: US Citizen
- Phone: (678) 900‑7101
I am a Computer Engineering student at the Georgia Institute of Technology with a 4.0 GPA, passionate about building innovative solutions at the intersection of artificial intelligence and software engineering. My practical experience includes developing full-stack web applications as an Engineering Intern at iVue Robotics and engineering a high-performance FastAPI backend to process real-time drone telemetry as an Undergraduate Researcher. I have a strong foundation in AI and machine learning, demonstrated by projects like 'Triage Assist,' an AI-powered system I built using Python and Scikit-learn , and my work fine-tuning neural networks with PyTorch to achieve over 90% accuracy in image classification. From architecting robot software in Java for the FIRST Robotics Competition World Championship to developing a complete Snake game in C++ on an mbed microcontroller, I enjoy tackling complex challenges and turning ideas into functional technology. .
Skills
Technologies I’ve used across projects, internships, and classwork.
Programming
Software
Hardware
Education
Georgia Institute of Technology
Graduation: May 2028
Bachelor of Science in Computer Engineering
- GPA: 4.0/4.0
- Relevant Courses:
- Intro to Artificial Intelligence — Gained a foundational understanding of AI by designing intelligent agents and systems using search algorithms, Markov decision processes, and neural networks.
- Design & Analysis of Algorithms — Mastered core techniques for the design and analysis of efficient algorithms, covering paradigms like divide-and-conquer, dynamic programming, and graph theory.
- Data Structures & Algorithms — Gained proficiency in programming and analyzing the performance of common data structures and algorithms, including various sorting, pattern matching, and graph traversal techniques.
- Programming HW/SW Systems — Explored the hardware/software interface by programming in C++, C, and RISC-V assembly, with a focus on systems-oriented design, performance, and memory considerations.
- Intro to Object Oriented Programming — Developed robust object-oriented solutions to computational problems in Java, covering core OOP principles, basic design patterns, data structures, and algorithms.
- Objects and Design — Worked in a team-based environment, utilizing Agile practices, version control (Git), and Django framework to develop a complex project
- Digital System Design — Gained hands-on experience in designing, implementing, and analyzing digital logic circuits using Boolean algebra, finite state machines, and modern design tools.
Georgia Institute of Technology
May 2025 - August 2025
Georgia Tech Europe Summer Study Abroad Program
- GPA: 4.0/4.0
- Countries Traveled: Germany, England, France, Italy, Greece, Portugal, Luxembourg, Croatia, Switzerland, Belgium, Netherlands, Austria, Hungary, and Czech Republic
Experience
Internships and leadership roles across web, IT, and robotics.
Software Development Engineering Intern
May 2026 – Present
Amazon Web Services
- Developing an AI‑powered agent using Amazon Bedrock to use weekly trend analysis and risk identification to autonomously propose and execute actionable remind/escalate actions with human‑in‑the‑loop approval
- Decomposing a large 1200‑line Lambda function into purpose‑built microservices, enabling independent scaling, deployment, and fault isolation and supporting real‑time analysis, while reducing operational costs
Data Analytics Intern, Technical Operations
January 2026 – May 2026
Delta Air Lines
- Engineered a machine learning pipeline utilizing KNN to label and categorize a dataset of 5,000+ service events; transformed raw historical invoice data into structured market insights for maintenance and repair benchmarking
- Developed application for engine bill‑of‑materials analysis, processing 10,000+ part relationships with hierarchical visualization and cost estimation capabilities in a multi‑tab GUI with dynamic part details
- Improved material cost model, estimating maintenance costs for engine modules based on workscope definitions
Meet Millennium
April 2026
Millennium Management
- Selected for an insight program exploring technology, trading, and research at Millennium.
- Participated in a competitive AmplifyME simulation, applying algorithmic logic and financial modeling to optimize portfolio performance in a simulated trading floor environment.
Engineering Intern
August 2023 – December 2025
iVue Robotics
- Engineered full‑stack web applications for iVue’s platforms (Vue.js, HTML, JavaScript, CSS) for web and mobile interfaces
- Integrated Cesium API to create custom data visualization layers for real‑time drone telemetry
- Facilitated $10,000+ of drone orders by developing a new product selection and orders page
Information Technology Intern
May 2024 – August 2024
Fulton County Schools
- Executed large‑scale reimaging initiative for 10,000+ Windows devices across 100+ schools
- Established proficiency in system imaging from PXE boot to final QA testing
- Collaborated within a team to troubleshoot and ensure proper operation for 100% of devices
Head of Business and Programming
August 2021 – May 2024
Storm Robotics
- Architected and deployed core operational software for FRC using Java
- Integrated sensor input and autonomous functionality for precise robot control
- Led onboarding on Java, Git, and hardware fundamentals; doubled the developer team in one season
- Qualified for FRC World Championship among 86,000+ students and 3,500+ teams globally
Advisor, Former President
August 2020 – Present
Inspiring Together (501c(3) Non‑profit)
- Organized 200+ weekly free online tutoring sessions for nationwide foster students with 100+ volunteers
- Led leadership team by assigning tasks and held weekly meetings to ensure consistent progress
- Managed $10k+ in funds and social media with 2k+ total following
Projects
Triage Assist
Built a scalable AI‑powered triage system integrating Llama 3.3 and a machine learning model trained using Random Forest, SVC, and KNN algorithms with 1200+ data points to assess and prioritize patients.
- Engineered a dynamic weighted ensemble technique to optimize model performance; fine‑tuned algorithm weights based on feature analysis and medical context, improving predictive accuracy for patient triage recommendations
- Designed a Streamlit‑based UI with dynamic color‑coded levels, handling for incomplete data, and a login system
Artificial Intelligence | CS 3600 @ Georgia Tech
Implemented 10 projects throughout semester spanning search algorithms, deep learning, neural networks (CNNs/LLMs), reinforcement learning (MDPs), and probabilistic inference.
- Engineered a quote search system using LLM text embeddings and FAISS for 500K quotes with Streamlit UI
- Fine‑tuned a pre‑trained VGG16 convolutional neural network (CNN) using transfer learning in PyTorch to perform image classification on a dataset of butterflies across 75 classes, achieving 90%+ accuracy
Image Classification using a Convolutional Neural Network
Trained a butterfly species classifier (75 classes) in PyTorch with ~90% validation accuracy via transfer learning.
- Fine‑tuned VGG16 by freezing early feature layers and training the final fully‑connected layers
- Built a prediction script to batch‑classify new images and export results to CSV
Predictive Model for Real Estate Pricing
Built a housing price predictor for Cleveland using 519 sales across five features (rooms, sqft, school proximity, etc.).
- Computed coefficients via both closed‑form matrix inversion and iterative gradient descent
- Achieved an R^2 score of 0.85, explaining 85% of price variability
Neural Network for Handwritten Digit Recognition
Implemented a neural network from scratch in NumPy to classify 28×28 MNIST digits.
- Built core components: fully‑connected layers, ReLU, softmax, and cross‑entropy
- Trained over 60 epochs with forward pass and backpropagation using gradient descent
Markov Decision Process Policy Optimization
Derived an optimal policy for a 12‑state MDP via Q‑iteration to maximize long‑term discounted rewards.
- Constructed and iteratively updated a Q‑table using probabilistic transitions, rewards, and discount factor
- Extracted the final greedy policy mapping each state to its best action
Bayesian Network for Probabilistic Modeling
Learned Bayesian network parameters from thousands of rows of data and performed probabilistic inference.
- Computed Conditional Probability Tables with add‑one smoothing to avoid zero‑probability issues
- Implemented a FullJoint class to evaluate joint probabilities for arbitrary variable assignments
Vertically Integrated Project - Experimental Flight
Developed Real-time Drone Telemetry API
- Helped develop a real-time drone monitoring system on Google Cloud Platform, creating a serverless pipeline (Pub/Sub, Cloud Functions, Firestore) for live telemetry and a scalable video streaming service.
- Engineered a high-performance, asynchronous backend API using FastAPI to ingest, process, and serve real-time telemetry data (e.g., GPS, attitude, battery) from a drone.
Mbed Snake Game
Fully functional Snake game using an ARM mbed microcontroller, uLCD display, and tactile switch input.
- Focused on real‑time input handling and smooth rendering
- Unit tests validated switch inputs, game logic, and display outputs