Robotics + Embedded Systems

Kethan
Vegunta

High school engineer working on autonomous underwater vehicles, RTOS firmware, and RF signal processing. Software lead at AVBotz, building the control, perception, and firmware stack for RoboSub; ML/AI intern at IBC.

Based
Bay Area, CA
01

Experience

Vice President & Software Lead

Aug 2025 — Present
Amador Valley Robotics Club · Pleasanton, CA
RoboSub Team — AUV Development
  • Led a 7-engineer software team across controls, firmware, and perception; coordinated with electrical and mechanical subteams via Interface Control Documents and directed the stack migration to ROS 2 Jazzy / Stonefish, contributing C++/OpenGL fixes upstream.
  • Replaced 6-axis PID with an LQR controller modeled in Simulink that linearizes underwater drag dynamics; rewrote thrust allocation as a QP optimizer that minimizes power draw and is tolerant to 2/8 thruster failures while still maintaining full 6-DOF control.
  • Wrote Zephyr RTOS firmware for sensor acquisition / actuation over CAN / UART; architected MAVLink comms between MCU and Jetson AGX Orin.
  • Trained YOLOv11 models and developed OpenCV post-processing algorithms in Python to detect competition props reliably across a variety of pool and lighting conditions; built Foxglove extensions used by the team for teleoperation and live model inspection.

ML/AI Intern

Nov 2024 — Present
International Business Consultants (IBC)
  • Architected ServiceNow components for Ethereal Agility: designed table schemas and data models; implemented Script Includes/Business Rules and Flow Designer/IntegrationHub flows; exposed prediction endpoints and role-based dashboards to product teams.
  • Built ALM data pipelines integrating Jira/Atlassian via REST APIs/webhooks: incremental ETL, schema mapping (sprints, issues, story points, assignees), and feature engineering (velocity, lead/cycle time, WIP, skills signals) joined with internal skills inventories for downstream modeling.
  • Developed timeline & allocation models using regression/time-series methods and constrained optimization to forecast delivery dates and recommend staffing; containerized/packaged as REST services, and integrated back into ServiceNow for in-workflow decisions.

Cluster 11: Brain-Inspired Machine Learning

Jul 2025 — Aug 2025
COSMOS · UC Santa Cruz
<7% acceptance rate
  • Studied cutting-edge techniques in the neuromorphic machine learning space including SNNs, memristors, and ternary quantization.
  • Applied these techniques to improve the performance of traditional machine learning models on classification and regression tasks.

Research Assistant

Dec 2024 — Jun 2025
San Jose State University · San Jose, CA
Assistant for a graduate student under Dr. Robert Chun
  • Engineered a RIASEC job recommendation system that fuses LLM-guided preference elicitation with implicit eye-tracking signals to rank career matches.
  • Built a data path from the Tobii C++ SDK: C++ publisher → ZeroMQ PUB/SUB → Python subscriber for feature extraction and scoring.
  • Developed a backend with a Flask server and socket-based communication with a Vite/React frontend.

Co-President

Jul 2025 — Present
ACE Coding · Pleasanton, CA
  • Conducted regular programming sessions at three middle schools: Hart Middle School, Harvest Park Middle School, and Pleasanton Middle School.
  • Oversaw operations at all branches and facilitated communication between Amador Valley High School and middle school administrations.
  • Organized ACE Code Day, an event with 150+ students, 20+ lecturers, and 14 workshops on software and engineering topics.

Vice President & Secretary

Aug 2024 — Present
Amador Valley Web Development · Pleasanton, CA
AV Web Dev
  • Presided over several projects and handled communication with the liaison from the organization or club that commissioned each project.
  • Made key technical decisions and contributed significantly to the code and design of commissioned projects.
  • Facilitated communication with school administration and completed the necessary paperwork.
02

Projects

Enhancing Disaster Rescue

Dec 2024 — Mar 2026
Localizing cellular devices via RF signal analysis
  • Prototyped a receive-only SDR localization stack (2× bladeRF 2.0 micro xA4 and 4 monopole antennas) to passively detect, analyze, and triangulate cellular RF emissions in visually obstructed environments for disaster-response use cases.
  • Wrote high-throughput C++ for DSP and fast IQ sample processing / acquisition; built custom GNU Radio blocks in Python to generate bearings at ~3 Hz.
  • Validated across diverse environments; achieved ~15° median bearing error and stress-tested performance under multipath and low-SNR conditions.
  • Won awards at ACSEF and CSEF — see the awards section.

Multimodal Pokedex

Jul 2025 — Aug 2025
  • Built an end-to-end pipeline — CNN image/audio classifiers → LLM-generated encyclopedia entries — across 6 classes and 11,000 total samples.
  • Shipped a Flutter app using ONNX Runtime for fully on-device iOS inference and deployed a web client via the Hugging Face Spaces API.
  • Analyzed the confusion matrix to optimize data collection, architecture choice, and training processes.

Table Tennis Game Annotator

Sep 2023
  • Trained a custom YOLOv5 model for ball detection and tracking; pruned and quantized the model for improved performance, achieving more than 10 FPS on the web.
  • Created a custom-trained Mask R-CNN model to segment the table, scoreboard, and players and provide live rally stats (rally length, score changes).

More at github.com/kethan1 ↗

03

Awards

ACSEF — 1st Place, Embedded Devices

2026 · Hayward, CA
  • Competed with Enhancing Disaster Rescue: Localizing Cellular Devices via RF Signal Analysis, winning 1st in the Embedded Devices category, the Stem4All Special Award, and the U.S. Air Force Special Award.

Lockheed Martin Code Quest — 1st Place, Advanced Division

2026 · Sunnyvale, CA
  • Won 1st among 30+ teams totalling 90+ people; first Sunnyvale team in history to score above 600 points.

USACO Platinum

Feb 2026
  • Promoted to the Platinum division with a Gold score of 667/1000 using techniques like dynamic programming, graph traversal, trees, and combinatorics.

CSEF — 3rd Place, Embedded Devices

2025 · Thousand Oaks, CA
  • Competed with Enhancing Disaster Rescue: Localizing Cellular Devices via RF Signal Analysis, winning 3rd in the Embedded Devices category.

ACSEF — 1st Place, Embedded Devices

2025 · Hayward, CA
  • Competed with Enhancing Disaster Rescue: Localizing Cellular Devices via RF Signal Analysis, winning 1st in the Embedded Devices category and the American Vacuum Society Special Award.

TSA National Conference

2026 · Washington, D.C.
  • 3rd Place, Coding — solved algorithmic problems in Python and C++ in an isolated room with no internet access.
  • 8th Place, Software Development — employed digital signal processing and ML models to localize and classify sounds for the auditorily impaired.

TSA National Conference

2025 · Nashville, TN
  • 5th Place, Coding — solved algorithmic problems in Python and C++ in an isolated room with no internet access.
  • 9th Place, Software Development — employed ML models to forecast future crop yields and potential improvements for farmers.

TSA National Conference — 3rd Place, Webmaster

2024 · Orlando, FL
  • Developed a website compiling information for homeowners on clean/green energy practices, initiatives, and tax rebate information.
04

Skills

Soft

Collaborative · Adaptable · Motivated · Detail-Oriented · Passionate · Reliable

Robotics & Autonomy

ROS 2 (DDS, QoS, lifecycle nodes, colcon) · tf2 · Nav2 · BehaviorTrees.CPP · Gazebo · RViz · rosbag2 · State estimation & sensor fusion (EKF/UKF) · Path planning (A*, RRT, Dijkstra) · Control (PID, LQR, MPPI) · Simulink

Perception & ML

PyTorch · scikit-learn · NumPy · OpenCV · ONNX Runtime · Camera calibration · Stereo & monocular depth · Quantization · Edge deployment (JetPack) · Dataset labeling (CVAT, Roboflow)

Embedded & Firmware

Zephyr RTOS · Interrupts & ISRs · DMA · Timers/PWM · ADC/DAC · Memory-mapped I/O · Watchdogs

Systems & Tooling

Docker · Git · GitHub Actions (CI/CD) · CMake · Make · pytest · Valgrind · systemd · netplan

Web Development

React · Next.js · Tailwind CSS · Flask · Node.js/Express · REST · GraphQL · OAuth2/JWT · PostgreSQL · MongoDB · SQLite · ORMs (Drizzle/Prisma)

Buses & Protocols

CAN · I2C · SPI · UART · OSI Model · Ethernet · TCP/UDP sockets · Multicast/DDS discovery · WebSockets

05

Education

Amador Valley High School

Graduating May 2027
PLTW Pathways: Computer Science · Pleasanton, CA
GPA
4.0 UW · 4.50 W
Relevant coursework
AP Calculus BCHonors Pre-CalculusAlgebra IIAP Computer Science ACybersecurityAdvanced Computer Science for the Contemporary WorldAP Physics C: MechanicsAP Physics C: Electricity & MagnetismAP ChemistryAP Biology