Skip to main content

Neeraj Gandhi

I like building systems that leverage co-design between hardware and software. Most recently, this has taken the form of working on Oracle Exadata Exascale, which provides high-performance cloud storage systems through tight integration of distributed systems software with the underlying hardware.

Prior to working at Oracle, I was researching how we can build safer and more secure distributed embedded / robotic systems as a part of my PhD at the University of Pennsylvania, advised by Professor Linh Thi Xuan Phan. I completed my PhD in 2024.

Previously, I founded and ran the NASA Robotic Mining Competition Team at the University of Virginia, with Professor Gavin Garner as our team advisor. I also was a part of the INERTIA Team at the University of Virginia, advised by Professor John Lach. I also had the opportunity to research in the PULSE Lab at Johns Hopkins University, with Professor Muyinatu Bell as my advisor.

Projects​

Most of my projects involve engineering. However, linguistics and languages are a hobby of mine, so I've done a few side projects in that space as well.

Technical Projects​

  • Second Silicon: A context-aware copilot for post-fabrication PCB bring-up, where schematics, firmware, and bench measurements meet.
  • Bounded-Time Interaction for Multi-Robot System Defense: Developing techniques to protect multi-robot systems from Byzantine faults with bounded interaction and recovery times.
  • NASA Robotic Mining Competition: Founding and leading a team (for four years) to annually develop a robot capable of mining regolith on Mars.
  • Ground-Up VLA Substack is my work at publicly building knowledge on how VLAs work. I have little background in machine learning, so this is my attempt to understand VLAs from first principles.
  • Bounded-Time Recovery: Developing techniques to apply Bounded-Time Recovery (BTR) to improve the safety of distributed cyber-physical systems that might experience either benign or adversarial (Byzantine) faults.
  • Rotor Fault Detection, Isolation, and Recovery: Detecting, isolating, and recovering from faulty rotors in modular aerial robots with dozens of rotors.
  • High-Performance Storage Systems: Over the past two years, I have gotten to dive into a large-scale, high-performance data storage system: Oracle Exadata Exascale, which achieves its performance via hardware-software codesign.
  • Photoacoustic Imaging for Surgical Guidance: Investigating how we can use photoacoustic imaging to reduce morbidity and mortality in surgeries (both traditional and robotic).
  • Dementia Agitation Detection: Designing PCBs, housings, and software for collecting sensor data and resident feedback in dementia patient homes.

Languages​

  • Phonetic Periodic Table of Elements: Proposing a systematic way to name chemical elements in Hindi and other Indic languages.
  • Navjani Script: Extending and adapting the historical Mahajani script into one with unambiguous spelling, distributed as an open-licensed font.

Skills​

Collectively, I have worked on various layers of a project stack, so I have experience in software design, PCB design, and mechanism design. I have worked on manufacturing these designs as well. Below is a selected list of the skills I've developed over the course of these projects, or out of interest.

  • Software/Programming: ROS, Gurobi, ns‑3 || C++, C, Python, MATLAB, Bash
  • Electrical/Electronics: Altium, KiCAD, Eagle, PIC, MSP430, XBEE, Raspberry Pi, Oscilloscope, Soldering
  • Mechanical/Manufacturing: SolidWorks, AutoCAD, CNC Mill, Water Jet, Plasma Cutter, Laser Cutter, Lathing, Mig/Tig Welding
  • Natural Languages: (Fluent) English, Hindi, Marwari; (Low Intermediate) Mandarin