NicoRamos

Since April 2025 · Yale Faboratory

Burrowing Soft Robot

I joined Yale's soft robotics lab in the spring of my junior year, working under PhD candidate Caitlin Le. The question is how a soft body's properties, like how stretchy it is or whether it has a rigid core, change the force that dirt or sand puts on it while it burrows. We measure that, then use the data to move toward a working burrowing robot.

• Python
• Mechanical design
• Data analysis
• Testing & troubleshooting

Senior capstone · Yale · team lead

Senior Design: Shingle Shark

Pulling up roof shingles is brutal work, and the tools that help are heavy pneumatic ones that need a compressor and stop being usable on steep roofs. I led a team of four to build something lighter and electromechanical: a compact, motor-driven tool that removes shingles and nails with less strain on the roofer. We built it module by module, and I led the motion generation and kinematics.

• Team leadership
• Kinematics
• Prototyping

Yale Exploration Rover · Drives subteam lead

Exploration Rover

I led a team of eight students designing the rocker-bogie suspension, propulsion system and chassis for Yale's rover for the University Rover Challenge 2025. We assembled the suspension and picked the motor and gearing to meet the obstacle requirement.

• 70 kg max load
• 1.2 m cube frame
• 15 mph top speed
• 0.5 m obstacles
• 26.9:1 gear ratio

• Onshape
• MATLAB
• Design for manufacturing
• Project management

Yale Liquid Rocketry · Propulsion subteam

Liquid Rocketry: Project Liquid

Our previous engine, Yengine 1.0, burned IPA and nitrous oxide. Yengine 2.0 is regeneratively and film cooled and runs on self-pressurizing ethane and nitrous oxide. I designed and machined the Augmented Spark Ignitor, which lights it.

• 0–3 lbf thrust
• 3.5 O/F
• 100 psi chamber pressure

• Onshape
• Lathe, mill, drill press
• Testing

Summer 2024 · Research

Autonomous Celestial Navigation for sUAS

Small drones under 55 lb do a lot of work in agriculture, disaster response and climate research, and most of them depend on GNSS, which can be jammed. I developed a daytime celestial navigation approach that doesn't rely on GPS or on features on the ground.

• Navigation
• sUAS
• GNSS-denied
• MATLAB

Jun–Aug 2025 · GlobalFoundries, Malta, NY

GlobalFoundries: Deuterium Recycling System

As an Equipment Engineering Intern, I worked on integrating a dual-redundant electrochemical hydrogen purifier. The goal was to cut about 99% of the money the company spends on deuterium each year and make it less exposed to price spikes.

• 99% target savings on deuterium

• Manufacturing
• Electrochemical hydrogen purifier

MENG 405 · Robotics Control

Franka Panda Pick-and-Place

For this class I wrote the whole control pipeline for a simulated Franka Panda arm. Inverse kinematics are solved numerically with a Jacobian pseudoinverse and an SE(3) error twist, motions use quintic time scaling for smooth acceleration, and a PD plus inverse dynamics controller uses the arm's live mass matrix to compute torques. The arm picks up a randomly placed cube and sets it at a random target.

• Inverse kinematics
• Trajectory generation
• Torque control

MENG 389 · Controls

Self-Balancing Motorcycle

We designed and simulated a self-driving motorcycle that balances with a reaction flywheel and moves with drive motors. It stayed stable with changes in payload and outside disturbances.

• 2× PID loops
• 9-axis IMU
• SAMD21 Arduino Nano 33 IoT

• MATLAB / Simulink
• Root locus
• Sensor integration

Personal project

Adaptive Drum Tuner

After ten years of playing drums, I can say tuning is hard. Every lug has to be brought up in small steps following a pattern, and doing that a few times per drum takes 10 to 20 minutes. I'm designing a tool that tunes all the lugs on a drum head at the same time.

• 10–20 min to tune one drum by hand

• CAD
• Product design

Astrophysics research

Estimating Black Hole Masses for Type 1 AGNs

An active galactic nucleus's mass comes from two numbers using the virial theorem: how far its broad line region is from the black hole, and how fast the gas there moves. The most rigorous way to get the distance, reverberation mapping, can take years for a single object, so I looked at faster ways to estimate it.

• ~5,000 AGN analyzed

• Astrophysics
• Data analysis
• Python