Mikul Rana · Saratoga High School, Class of 2027
I'm a high-school engineer focused on assistive and adaptive robotics — devices that give people back abilities their bodies have lost. My first real project was a hand exoskeleton for my grandmother, who has Parkinson's; I've since carried that work into a university research lab and a nonprofit I founded.
What ties everything together is a single idea: the machine should adapt to the person, not force the person to adapt to the machine. Whether it's a glove, a dog's wheelchair, or a research rig, I care most about the moment a real user gets a piece of their independence back.
Solo garage prototypes — a hand exoskeleton built for my grandmother, iterated through three failures.
Research at the SFSU CARE Lab — learning to measure, log, and prove what "works" actually means.
Bringing engineering to real community accessibility needs — designing with users, not just for them.
A 468 g tendon-driven wearable that restores grip for someone losing hand dexterity — with YOLO-based adaptive selection between pinch, tripod, and power grips. Engineered across three iterations, each one a response to a real failure.
The problem. My grandmother's Parkinson's made everyday tasks — lifting a spoon, holding a glass — a fight against her own tremors. Commercial hand exoskeletons ran from $800 into the thousands, and none were designed around her needs. So I built one.
Fishing line routed through 3D-printed rings on a glove, pulled by servo motors driven by EMG muscle sensors.
✗ Failed — tremors made EMG signals inconsistentA voice-recognition module — she says "grip," the fingers close around the object. She could drink on her own again.
✗ Worked, then failed — Parkinson's took her voiceAn onboard camera + CV model recognizes the object and picks the grip automatically — five fingers around a glass, three to pinch a spoon.
✓ She could eat on her own — the device adapts to herThe real shift wasn't technical. Each version stopped asking my grandmother to change herself for the device, and started making the device change for her.
A faculty-mentored continuation of my exoskeleton work — where I learned to investigate rigorously: measure, log, benchmark, and prove.
Python tools to log, visualize, and sync EMG muscle data with recorded video (CSV/TXT output).
Timestamped video, trial segmentation, metadata, and live researcher annotation for repeatable human-subject trials.
STM32 Nucleo + Analog Discovery 2; raw-binary transmission at 691200 baud → ~1.9 ms latency.
CAD'd a Raspberry Pi 5 camera mount; traced a bounding-box failure to bad model weights; added a keyboard fallback.
CAD and resin-printed components for camera integration and the experimental platform.
Prepared and ran trials, operated data-collection tools, and gathered synchronized multimodal data.
Coauthor on a research paper targeted for submission to IEEE-RAS Humanoids 2026.
7 dogs helped
Adjustable dog wheelchairs built from aluminum/PVC frames with custom 3D-printed joints, restoring mobility for injured and aging dogs. Started when my own dog needed one after TPLO surgery; each build is iterated with the owner and veterinary contacts.
A convolutional neural network that flags early Parkinson's motor decline from hand-drawn spiral patterns — the same disease that motivated the exoskeleton, approached from the software side.
Manage hardware, manufacturing, and electronics inventory and lab safety for a competitive robotics team; former hardware subsystem lead. Mentor students in CAD, machining, and assembly.
Precision CAD work on parachute components for an aerospace recovery-systems company — real manufacturing tolerances, real parts.
Teach beginner and intermediate CAD to middle- and high-school students, and mechanics to VEX Robotics teams — turning the tools I learned building into something I can hand down.
Lead bridge-prototype design and physics instruction, and captain the TSA TEAMS squad — twice qualifying for the national competition.
Open to research collaboration, assistive-technology projects, and conversations with labs and mentors. The fastest way to reach me is email.