Project Overview

Gumps is a modular, magnet-based toy designed to empower kids with limited hand dexterity to build, play, and express themselves creatively, without the frustrations caused by traditional fine-motor-heavy toys.

Built with a 'design for one, extend to many' mindset, Gumps is accessible, joyful, and endlessly customizable, inviting all children to create their own monster characters with soft, safe, easy-to-use magnetic parts.

• Age Range: 5–7 years
• Materials: Soft body + magnetic attachments
• Core Values: Accessibility, Creativity, Replayability

The Problem

Many popular creative toys, like LEGOS, rely heavily on fine motor skills, intricate manipulations, and tight fittings, creating barriers for children with hand impairments.

Gumps was born to eliminate those barriers, fostering an inclusive, frustration-free world of creative play.

The Process

Research and Insights

Interviews with childcare professionals

Competitive analysis with product positioning matrix

Market trend analysis (inclusive toy market growth)

User interviews with educators

Competitive market analysis

User persona development

Strategic opportunity mapping

Early Ideation and Prototyping

Explored magnetic tables and block systems

Identified challenges like separation force and user frustration

Initial concept sketches

First functional prototype

"I like to start ideation with a diverging exercise, where I explore many possibilities. My team and I independently came up with as many ideas as posisble following our research. Then, we reconvened and converged on the ones that fit our design criteria the best."

Rapid Iteration

Shifted to soft-bodied monsters with magnetic attachments

Focused on tactile feedback and easy customization

Second iteration with improved magnets

Final prototype before production

"Design (inclusive design especially) is a conversation. Based on our feedback from the first round of prototyping, we decided to pivot to a larger format toy. This allowed us to focus on the core functionality and user experience, while still allowing for a high degree of customization."

Modeling and Testing

Built physical prototypes and tested for usability

Optimized material softness, magnet strength, and modularity

3D modeling for final design

Final production-ready design

"It is hard to tell if an idea if achieving its aims until a user can test it. With our prototypes finalized and validated, I was able to model the first Gump, "Splat", and build him for kids to test."

Impact and Reflection

Gumps proved that inclusive design choices — from softer materials to intuitive feedback — can create toys that are better for everyone, not just specific audiences.

It became a joyful experience not only for kids with dexterity challenges, but for all kids who love building, swapping, and creating.

Designing Gumps deepened my belief that true innovation comes from focusing on the edges — solving for specific needs while opening new possibilities for all users.

It's a lesson I carry into every project: build with empathy, design with courage, and create experiences that invite everyone in.

Project Overview

Selective Laser Sintering (SLS) 3D printing unlocks incredible possibilities for material science, biomedical research, and custom manufacturing.

But commercial SLS printers remain prohibitively expensive, proprietary, and inflexible — locking out researchers who need open, tunable systems.

This project reimagines SLS accessibility by creating an open-source optical control system — enabling precision laser scanning using low-cost hardware and custom electronics.

The Problem

Traditional SLS printers rely on tightly guarded optical control systems to guide their lasers with micron-level precision.

Without affordable, open alternatives, researchers are forced to work around expensive black-box technology — limiting innovation in critical fields like material development and biomedical engineering.

The Process

Research and Setup

Reviewed commercial designs (e.g., Adafruit Feather) to understand core functional requirements.

Learned in-house PCB fabrication processes, including laser etching and manual ultra-small assembly.

"The journey to miniaturization started with understanding the fundamentals. By studying commercial designs and mastering fabrication techniques, we built a foundation for pushing the boundaries of what's possible at the microscale."

First Iteration: Conceptual Framework

Designed initial PCB and firmware to translate g-code step pulses into DAC outputs for galvanometers.

Discovered IO and signal integrity limitations in early tests.

Schematic for v1

Layout for v1

Render for v1

Live board v1

Second Iteration: Hardware and Signal Refinement

Redesigned PCB with improved op-amp signal paths and thumbwheel potentiometers for tuning.

Added extensive test points.

Improved but encountered analog noise issues at high scan speeds.

Schematic for v2

Layout for v2

Live board v2

Third Iteration: Precision Tuning and Integration

PCB revision achieved clean differential analog outputs with full galvanometer control.

Verified 6m/s laser scan speed and minimum feature size of 0.018mm.

Validated full g-code integration with 3D printer software.

Schematic for v3

Layout for v3

Render for v3

Fourth Iteration: Final Integration and Testing

Refined the PCB design for ease of use and integration with 3D printer software.

Improved test points and documentation.

Upgraded potentiometers for better resolution and tuning.

Schematic for v4

Layout for v4

Render for v4

Impact and Reflection

Making SLS printing more accessible unlocks entirely new possibilities for research and innovation.

This development has gotten us one step closer to the prototyping of custom 3D-printed bone scaffolds — showing the potential for personalized biomedical models that were previously out of reach.

I was pushed to the edge of my technical abilities through this project — across circuit design, signal processing, firmware architecture, and iterative debugging.

It taught me that the hardest technical challenges are often the most rewarding — and that persistence, precision, and curiosity are what transform impossible-seeming problems into real-world solutions.

Project Overview

Nanoscale robotics holds transformative potential for medicine — from targeted therapies to micro-surgical interventions.

This project focused on developing an ultra-miniaturized wireless control system: a custom PCB housing an nRF52832 and magnetic sensor, small enough to fit inside a medical stent and control magnetically actuated micro-robots in vitro.

The Problem

Traditional PCB architectures are too large, power-hungry, and rigid for nanoscale medical applications.

To enable real-world micro-robotics, we needed a functional, hand-assembled PCB no larger than 8mm — combining wireless communication, magnetic sensing, and robust manufacturability at extreme scales.

The Process

Research and Setup

Reviewed commercial designs (e.g., Adafruit Feather) to understand core functional requirements.

Learned in-house PCB fabrication processes, including laser etching and manual ultra-small assembly.

"This journey was an exciting opportunity to learn about in house manufacturing and design on a miniature scale. I began the process by learning and practicing the fundamentals of nanoscale work and in house PCB manufacturing. I then embarked on a year's long iterative process."

Iteration 1: Scaling Up to Scale Down

Built a large (~1 inch) replica of the Adafruit nRF52832 board.

Used as a platform to master programming, firmware flashing, and laser fabrication.

Schematic for v1

Layout for v1

"Starting with a larger board was crucial for understanding the fundamentals. It allowed me to experiment freely, master the manufacturing process, and build confidence before tackling the challenges of miniaturization."

Iteration 2: First Miniaturized Attempt

Designed an 8mm board using a bare nRF52832 chip.

Required an external antenna, limiting compactness and introducing mechanical weaknesses.

Schematic for v2

Layout for v2

Live board v2

Iteration 3: Integrated Antenna Experiment

Switched to a packaged nRF52832 with an integrated antenna.

Placed the chip centrally to save space — but learned that antenna placement must be at the board's edge for functional wireless communication.

Layout for v3

Live board v3

"This iteration was a crucial step in understanding the tradeoffs between antenna placement and board size. It highlighted the importance of careful design for even the smallest systems. It was also the point where scale really started becoming a challenge. For reference, the image here is about the size of your fingernail, and the wires were as this as hairs. I had to develop precise strategies to further manipulate and test this board."

Iteration 4: First Fully Functional Design

Redesigned layout with edge-mounted antenna.

Achieved a 6mm board with working communication and magnetic sensing.

Bulkier square footprint; functional but not optimal.

Layout for v4

Bare board v4

Soldered board v4

Iteration 5: Final Optimized Form

Refined the design to an 8mm elongated board — the smallest feasible footprint with the tools available.

Balanced antenna function, mechanical strength, and manufacturability.

Final boards successfully integrated wireless control and magnetic field sensing for micro-robot actuation.

Layout for v5

Live board v5

"This iteration was the culmination of all the previous work. It was the most challenging iteration, but also the most rewarding. I was able to achieve a functional board that was small enough to fit inside a medical stent. By this point, I was confident in using the microscope as my better pair of eyes and etching new PCBs was almost a passive endeavor. This form was finally testable in the true stent this project was meant to serve."

Impact and Reflection

This project taught me that engineering at the edge of manufacturability demands not just technical skill, but resilience, creativity, and obsession with detail.

Through each iteration — from large-scale copies to sub-centimeter functional systems — I gained a deep appreciation for the art and science of precision engineering.

True innovation often hides in the smallest details.

Project Overview

Our population is aging rapidly — and with it, the burden of memory loss and dementia grows.

Mora is a Personal Memory Language Model™ (PMLM) that uses AI and AR glasses to support daily memory, identity, routines, and connections — helping people maintain independence longer as they age.

This project focused on strategic opportunity framing, user validation, and early venture development for Mora.

The Problem

1 in 4 people will be over 65 by 2050 — and memory-related conditions are surging globally.

Yet most existing memory support technologies are reactive, clinical, and disconnected from daily life.

There is an urgent need for human-centered, proactive cognitive support solutions that integrate seamlessly into people's lives.

The Process

Strategic Opportunity Identification

We began by studying Pattern Breakers by Mike Maples Jr., which challenged us to think not just about trends, but about major inflections — permanent shifts that change the future landscape.

After evaluating potential intersections, we focused on the convergence of AI and aging as the most powerful and urgent forces shaping the future.

Drawing on insights from my work with 7D and deep experience in the disability and aging space, I understood a critical truth: independence is everything for aging populations.

From this strategic foundation, the concept of a personal memory assistant emerged — a system that could evolve with users over time, helping them retain their memories, routines, and human connections.

This vision ultimately evolved into Mora.

Timing analysis for market entry

Competitive landscape analysis

Strategic opportunity matrix

Market and Business Development

Mapped total addressable market for elders with tech access and memory support needs.

Analyzed competitive landscape across dementia support tools, AI assistants, and AR wearables.

Positioned Mora uniquely at the intersection of memory health and everyday life.

Created early business model canvases and financial pro formas.

Developed go-to-market strategies through caregiving organizations, healthcare pilots, and direct-to-consumer channels.

Total addressable market analysis

Business model canvas

Financial projections

Market Research Doc

Product Ideation

Building on our strategic insights, we began exploring how AI and AR could work together to support memory and cognition.

The winning idea was a personal memory language model that uses AI and AR glasses to support daily memory, identity, routines, and connections.

Mora is an app that integrates into existing and future systems, and ages with you as time goes on. It sees and feels your entire life, living it with you, and keeping that data ready for when you need it most.

This ideation phase helped us define Mora's unique value proposition: a system that evolves with users, providing gentle support early and deeper cognitive augmentation as needs grow.

Initial product concept

AR interface design

User interaction flow

Key use case scenario

User Research and Validation

We then conducted parent test interviews with dementia patients, caregivers, and experts.

Some key insights we learned were that patients with dementia often suffer from paranoia, and are highly isolated.

Ultimately, the product need validated through both direct interviews and expert outreach.

Personal Memory Language Model architecture

User research insights

Parent Test

Impact and Reflection

Strategizing Mora reflected and reinforced my belief that real innovation happens not just by creating new technologies — but by deeply understanding human needs, fears, hopes, and futures.

Strategic innovation demands as much empathy as it does ingenuity.

Mora reflects a future where memory loss is met not with fear, but with empowerment.