About Me

Darshit Desai

Darshit Desai


Robotics Engineer

University of Maryland, College Park | Freefly Systems


Multidisciplinary Robotics Engineer with expertise across perception systems, embedded software, and aerial robotics. Experienced in full-stack robotics development—from firmware and hardware integration to AI-based sensing and autonomy. Proven success in automation, UAV systems, and applied research environments.


Work Experience

Robotics Engineer

Freefly Systems Inc.

June 2024 - Nov 2025

Freefly FLUX LiDAR:

  • Led firmware development for the Freefly FLUX LiDAR to enable precise, time-synchronized multi-sensor data capture — designed and implemented LiDAR, IMU, and GNSS drivers with a shared-memory-based logging system.
  • Developed applications on Zynq Ultrascale MPSoC, optimizing user space applications and kernel space drivers for critical timing operations.
  • Developed a 20 MP RAW Ximea camera driver at 5 FPS for perception tasks including pointcloud colorization and VIO.
  • Designed the factory target-based LiDAR–camera calibration workflow and extended it with a targetless field calibration feature for customer use.
  • Developed ROS 2 driver hooks for all onboard sensors using a timesync framework for developer efficiency and prototyping.

Dual-Operator ROS 2 Gimbal Manager:

  • Ported and redesigned the gimbal manager from STM controller to SoM running ROS 2, enabling dual-operator aerial camera control for high-end cinematography.
  • Delivered the system to beta customers for FIFA World Cup 2026, meeting low-latency performance targets and robustness criteria.
  • Strengthened system security by implementing ROS 2 microDDS authentication and access control using public/private key infrastructure.

Freefly Manufacturing Automation:

  • Developed a multi-bringup software pipeline for the Skynode flight controller, enabling simultaneous flashing of the SoM and STM32 MCU, reducing operator steps and errors.
  • Designed an automated bringup and testing station to validate onboard sensors and I/O pins; introduced a Make-based parallel testing architecture that reduced validation time from 20 minutes to 6 minutes.
  • Increased production throughput by creating an automated flashing system capable of programming 20 GNSS modules concurrently.

Motor Drive Firmware Development & Testing:

  • Designed and validated Hobbywing X9 propulsion system test setups, leveraging CAN and UART for synchronized telemetry logging and fault injection across multiple motor drives.
  • Automated FOC firmware stress testing with Python scripts simulating PX4-style slew rate constraints, generating new PWM sequences every 25 runs for dynamic test coverage.
  • Performed tuning of the sensorless FOC drive and improved motor efficiency from 85 % in Astro Prime to 92 % in Astro Max motor propeller pairings.
  • Designed state of the art sensorless drive firmware for the Astro MAX motor drives. Complete system design from STM32 bootloaders to app firmware.>/li>
  • Tuning of Sensorless FOC drive, with custom python tools for testing like 1 kHz CANbus telemetry logging and drive actuation commands. Automated noise injection, cycle testing and lifecycle stress testing of the propeller power plant.

Teaching Assistant, Matrix Air Land Sea Lab

University of Maryland

June 2023 - May 2024
  • Developed motion tracking system which tracks 20+ drones using motion capture cameras with reflective markers and Kalman filtering.
  • Designed a communications setup for reliable low-latency data transfer for 20+ drones over Bluetooth Low Energy radios.
  • Developed C++ code for performing group operations like takeoff and landing using VIO poses for localization.
  • Developed a drone-to-drone tracking system leveraging a re-trained YOLOv5 model to identify neighboring drones in the swarm.
  • Developed a sensor fusion algorithm fusing object detection and point cloud data at 30 FPS real-time speed.
  • Designed a tracking system using C++ and TensorFlow Lite libraries leveraging the Adreno GPU onboard the drone.
  • Implemented CV-based Kalman filter using Eigen3 for predicting acceleration and velocity of tracked drones.
  • Project goal: Deploy a decentralized flock of swarms leveraging poses calculated by fusion of RGB and Time-of-Flight cameras.

Graduate Student Researcher

Daikin Energy Innovation Lab, CEEE, University of Maryland

December 2022 - May 2024
  • Designed 3D-printed nozzle in SolidWorks with 2-degrees of freedom actuated by servo motors for HVAC system.
  • Developed tracking algorithm using thermal camera to dynamically reorient the nozzle using temperature segmentation.
  • Upgraded existing tracking system cameras to stereo cameras with sensor fusion of depth and RGB data.
  • Integrated YOLOv8n detection algorithm for object detection in RGB channel for accurate spatial estimation.
  • Integrated and repurposed Kalman filtering for tracking, improving spatial accuracy by up to 3x compared to previous versions.
  • Designed circuit boards and power distribution boards for the tracking system, integrating Raspberry Pi and Jetson Nano.
  • Designed PCBs and voltage regulators for DC fans and motors with feedback-based control for nozzle positioning.
  • The designed thermal system is estimated to provide 12% to 30% energy savings compared with typical conditioned building systems.

Assistant Manager, Press Manufacturing Engineering Division

Suzuki Motor Gujarat Private Limited

July 2018 - Aug 2022

Projects Undertaken at SMG:

  • New Product development:
    • Supported in achieving production targets by identifying machine related modifications and completing it before the start of vehicle’s mass production.
    • Design of material handling skids, pallets and jigs for New Vehicle Model die tooling. Acted as a point person for standardizing designs of various material handling equipment.
    • Supported ramp up phase of Baleno, Swift and Dzire models (1000 vehicles/day production volume models) with timely delivery of the required skids, pallets and fixtures.
    • Ensured customer feedback after handover is implemented on to the next car model preparation by revising standards and maintaining issue tracking lists.
  • IOT System:
    • Developing a SCADA based system for collecting quality related data from Press Machine with the end goal of integrating industry 4.0 features.
    • Analyzed data collected from the IOT system to establish Process control methods like Statistical process control to generate system based alarms for defective parts.
  • Crack Monitoring System:
    • Implemented a vision based sheet metal crack detection system where 20 cameras were used to detect surface defects on sheet metal car components.
    • Performed user acceptance testing of the software and hardware to recommend further development and modifications in image recognition, capturing and storing codes.
  • Integration of Crack Monitoring System and Press machine:
    • Applied MODBUS protocol for integration of the Press machine with the above system.
    • Programmed Press machine Digital I/Os for development of part recipe change management software and successfully tested synchronization with Press machine.
  • New Facility Establishment:
    • Instrumental in establishing a green field plant of 250,000 vehicles/year.
    • Handled 50+ vendors working under the same roof for Installation & commissioning of Servo Press Machines and other allied equipment of New Stamping Press Shops.
    • Streamlined progress with internal stakeholders like the Civil and Utility team to achieve the target of commissioning as per the timeline.
  • Particle Counter System:
    • Installed Laser based particle counters at various locations and conceptualized the PLC network design.
    • Assessed particle count and to establish clean rooms in areas of high importance and high contamination inside the Press shop to achieve defect rate reduction.
  • Stamping Press Final Goods Traceability system:
    • Devised architecture and flow of a system which integrated the vehicle tracking data like VIN with Press shop raw material data into a Centralized repository for back tracing.
    • Coordinated with 3 different shop production teams in getting feedback during design and benchmarking of the system.
  • Promotions: Joined as Graduate Engineer Trainee in 2018, Promoted to Junior Manager in 2019. Promoted for the second time as Assistant Manager in 2020.

Educational Lineage


Masters in Robotics

University of Maryland

2022 - 2024

Worked on Autonomous aerial robotics, swarm robotics, object tracking and 3D Vision.



Bachelors of Technology in Electrical Engineering

Nirma University, Ahmedabad, India

2014 - 2018