Dhillon Patel

01 · About

A bit about me

I'm a software engineer on the MX Routing team at Cisco Meraki, where I work on the networking stack that powers thousands of security appliances — from VPN NAT logic to ECMP and BGP-based dynamic routing. Lately I've been leading our team's AI initiative: architecting a multi-agent LLM workflow that takes a support ticket end-to-end to a merge request, using a graph-based approach to orchestrate planning, code generation, and review agents.

Before that, I built embedded firmware at Nordson, configuring STM32 microcontrollers, JTAG, and Ethernet PHYs, and got my start at Meraki as a firmware intern overhauling test infrastructure with Jenkins, Gerrit, and Yocto.

I graduated Summa Cum Laude from the University of Michigan with a B.S. in Computer Engineering, where my projects spanned computer vision, robotics, and embedded control systems.

In my free time I like to play basketball, tennis, and golf, snowboard, and watch football and F1. Lately I've been learning latte art on my new espresso machine.

Roles across networking, embedded & AI

500+

Support engineers using tools I've shipped

02 · Experience

Where I've worked

From routing infrastructure to embedded firmware to agentic AI tooling.

Mar 2025 — Present

CURRENT

Software Engineer, MX Routing Team

Cisco Meraki · San Francisco, CA

Leading the team's AI initiative, architecting a multi-agent LLM workflow that takes a ticket end-to-end to a merge request, using a graph-based engineering approach to orchestrate planning, code generation, and review agents
Design and maintain VPN NAT logic supporting 1:many and many:many address mappings for MX Security Appliances, conserving IP address space across large-scale customer deployments
Develop and extend ECMP (Equal-Cost Multi-Path) routing and BIRD-based dynamic routing to improve traffic distribution, failover behavior, and route table reliability
Overhauled the route table page by introducing protobuf-based firmware-side filtering, drastically reducing page load times for large route tables
Built a Linux ip-style wrapper script for commonly forgotten table read handlers, saving roughly 10 seconds per lookup for the ~500-person support team who run these dozens of times a day

Oct 2024 — Feb 2025

Embedded Software Engineer

Nordson · Amherst, OH

Developed embedded software for STM32 microcontrollers, configuring JTAG and pin configurations, and integrating an Ethernet PHY and Ethernet switch chip
Designed a command-line interface for the bootloader, reducing software download time by 20 seconds compared to the GUI version

Sep 2023 — Dec 2023

Firmware Software Engineer Intern

Cisco Meraki · San Francisco, CA

Designed and integrated new testing pipelines using Jenkins and Gerrit, accelerating the team's development process
Added code coverage to native test recipes, improving test effectiveness by 30% through CMake, Makefiles, Autotools, Ruby scripts, and BitBake recipes in Yocto

03 · Projects

My Projects

Academic and personal projects spanning computer vision, robotics, controls, and embedded systems.

RGB-D Semantic Segmentation

Trained multiple semantic segmentation models, demonstrating improved accuracy with the NYUv2 dataset after adding a depth channel in Python with the PyTorch library.

PythonPyTorchComputer Vision

↳ View on GitHub

Concussion Detector

Researched and developed a wearable IMU-based neck/head sensor for better prediction and diagnosis of concussion/whiplash injuries in athletes, using STM32 microcontrollers, UART, I2C, SPI, and ported drivers in C.

STM32CUART/I2C/SPI

↳ View on GitHub

Autonomous Robotics — Pursuer-Evader

Implemented pursuit-evasion behaviors for two differential-wheeled robots using Python and C++, integrating LCM and OpenCV for real-time audio and visual tracking, achieving autonomous navigation and target acquisition in dynamic, cluttered environments.

PythonC++LCMOpenCV

↳ View on GitHub

Adaptive Cruise Control

Developed a Simulink model for a vehicle with Adaptive Cruise Control and haptic steering, where wheel position controls steering angles and feedback forces. Implemented differential equations for vehicle dynamics and a PID-based automatic steering controller, interfacing with ADC, CAN bus, and digital inputs.

SimulinkPID ControlCAN Bus

↳ View on GitHub

Robot Kinematics and Motion Planning

Projects in JavaScript and HTML5 covering robot middleware and pub-sub messaging, graph-search path planning (A*), physical simulation (Lagrangian dynamics, numerical integrators), forward/inverse kinematics (matrix transforms, quaternions, Jacobian methods), and sampling-based motion planning.

JavaScriptA* SearchKinematics

↳ View live demo

05 · Education

Academic background

University of Michigan

B.S. Computer Engineering · Ann Arbor, MI

Aug 2020 — May 2024

Relevant courses: Computer Vision, Embedded Control Systems, Autonomous Robotics Design Experience

Honors & Awards

Summa Cum Laude James B. Angell Scholar Constance B. Mathias Engineering Scholarship Richard Earhart Scholarship Dean's Honor List University Honors William J. Branstrom Freshman Prize

3.87

GPA / 4.00

A bit about me

Where I've worked

My Projects

Technical Skills

Academic background

Let's build something reliable.