For many uni projects and during free time, I like to make renders of concepts, real products or other interesting scenes.

I was inspired by professionals such as Blender Guru (YouTube) and @sam_does_design, who produce incredibly beautiful renders that both convey a product idea and are satisfying to look at.

My preferred renderers are Blender and Solidworks Visualize.
Blender offers unparalleled versatility and a wide community for teaching and support.
SW Visualize is a quick and easy way to turn Solidworks models into beautiful renders.
I am familiar with Keyshot but do not use it due to it’s lack of a GPU rendering engine.

As my main second year project I developed a wearable device that uses haptics to guide a visually impaired user in a city environment.

This was a group project, we split the workload into roles, I chose to be the CTO. Also in my group were Angela Ibanez, Mia Taicher and Elvire Coudray.

The project involved developing the electronics back-end for the bracelet, circuit design, and the iOS app that interfaced with it.

For three months we researched and conducted interviews. We discovered that the most dangerous part of a user’s day was the journey, particularly when navigating alone.

To try and solve this problem we decided to eliminate voice instructions completely and use discreet haptics to send turn by turn directions to the user.

The bracelet was designed to interface with an iPhone app. We went with the apple platform because it offers far more accessibility features than any other smartphone and hence the VI market predominantly uses apple products.

To bypass the need for MFI certification, I developed a custom circuit that will amplify analogue sound signals. I then purchased a small bluetooth receiver and connected it to the phone just like a pair of headphones.

For the phone side of the project, I learned Swift and wrote a navigational app that will extrapolate turn by turn directions, and when the user breaks a waypoint geofence it will stream a specific audio haptic recording to the bracelet. If the user started moving in the wrong direction the app would also react and stream a ‘Check again‘ haptic to the bracelet.

The final bracelet was designed to be fully injection mouldable and contained seven components: two bluetooth boards (one for each channel), two haptic motors, the custom PCB, a button and a Li-Po cell.

My physical computing project involved designing, fabricating and testing a mechanical electronic device.

The scope was very open with the only requirement being a human interaction.

The goal was to teach us about sensors, actuators and control systems.

My idea involves a large dial that the user could rotate.
It had to be in the correct position to proceed to the next level.

An indication is given to the user by the brightness of an LED - the brighter it is the closer they are to winning.

To make the game more challenging I added a powerful stepper motor connected to the disk through a gear train.

If the user made an incorrect move or spun the disk too slowly the motor would add resistance to the motion, or completely take over fighting the user for control.

The game is beaten by either not making any mistakes (the patterns can be learned), or by going with brute force and overpowering the motor.

During my first year at Imperial College I worked on a project aiming to promote health and well being.

My team focused on improving the portability of commonly used sports equipment through prototyping and human centred design.

The team consisted of myself, Josh Moody, Angela Ibanez and Rachel Brown.

My proposal consisted of a collapsible badminton racket that would easily fit in user’s bag without taking up much space.

After conducting all the necessary research and some human factors measurements, I produced low fidelity blue foam prototypes, testing the experience.

Once I was confident about the concept, I moved on to high fidelity and functional prototypes which I later combined into a functional product.

The racket has an aluminium design with titanium and nylon composite connectors that ensure the racket will not come apart during heavy use.

During summer 2017 I decided to attempt making my own power bank.

The goal was to make something universal and have a very large capacity, rarely needing to be charged.

I tried to find complete modules and assemblies where possible since at the time my manufacturing resources were limited to a soldering iron and hot glue gun.

My second year design project explored the environmental impact of FMCG products.

After tearing down donor products, analysing their cradle to grave journey and understanding the product service system as a whole, we developed a proposal for a more sustainable solution.

In summer 2016, following the rumours of an iPhone sans audio jack, I built a prototype dongle.

Simple audio transmission would not suffice, it also had to integrate in-line remote control and mic functionality as those features now come with most headphones.

I implemented the remote detection by adding an arduino controller that could monitor is anything was pressed or when to turn on the mic.

To keep it simple to use I used the AUX port as a power button: it would turn on once headphones were inserted.

Unfortunately due to the difficult nature of in-line remote headphones coupled with the small form factor requirement, this project did not live for very long. But it certainly was a great learning experience!

During my Emotech internship I became interested in mechanical keyboards and the DIY community. To further my knowledge I decided to built my own keypad from scratch.

The most challenging part was designing the PCB and deciding on a microcontroller.
I chose the Atmel Mega32u4 running the Arduino bootloader as I was already familiar with the language.

I created the PCB layout in Eagle and ordered a Chinese manufacturer to make them. (I considered making them myself but ordering was cheaper and less hazardous)

Since this was to be a macro keypad, I programmed the Atmega to inject common keystrokes rather than individual characters.

Overall the project was a great success and worked first time, I am currently prototyping custom keycaps for it using my DLP printer.

As part of my engineering studies I learned to use Computer Aided Engineering (Finite Element Analysis) to simulate structural events on model parts. This helped me determine how strong a part would have to be made (and with which material) in order to satisfy the design requirements and have an acceptable safety factor.

The process involved modeling a part in solidworks, generating a object mesh and stimulating or with unit or cyclic loads.
An important process was iterative mesh refinement. A draft mesh will show which areas have the most stress, then those areas will have a finer mesh to make more accurate calculations.

As part of my coursework I had to iteratively design a bike frame, using FEA to determine the advantages and drawbacks of each design and improving on it.

For another coursework I was in a group with Will Kerr, Aida Manzano and Lauren Dowling.

We developed a water tower to be used in sub Saharan communities.

For this project the cost of the tower had to be minimized so the design had to be as simple as possible in order to balance performance with cost.

Once the model was ready we built miniaturized our design and built it out of bucatini pasta to build it in real life.

Bucatini was chosen over spaghetti since it has a higher second moment of area and hence is more strong as well as more delicious on a plate.

As part of my second year, I did a module in ML, Statistical Analysis and Big Data.

The main task involved analysing data and running appropriate t-tests to determine if a certain factor was significally different across multiple controlled tests. In my case I was determining how different manufacturers of different capacity memory chips could withstand different amounts of program erase cycles before failure.

To consolidate the module I did a group project with Will Kerr and Rob Garland. We analyzed 8000 instances of speed dating and determined the key variables that lead to a successful date with an accuracy of ~75% (Precision: x and Recall: y)

This module taught me the key principles behind machine learning. I am planning to do some personal projects in this area to expand my knowledge and satisfy my interest.