about
My name is Defne Dilbaz.
I am a 25 year old Software Engineer at Intel Corporation. I graduated from the University of Toronto Electrical and Computer Engineering in 2022 with minors in Engineering Business, and Robotics and Mechatronics; and a certificate in Artificial Intelligence Engineering.
During my undergrad, I focused on computer hardware and control streams. During second year, I built a pattern memory game: in Verilog with my Digital Systems course partner, Sarhad Salam. For ECE297, Design and Communication course, I developed SquadMaps - a geographical information system (like Google Maps) in C++ - with Sarhad Salam and Zeng Zeng. In third year, I took part in Bridgestone World Solar Challenge 2019 as an electrical-electromechanical team member for Blue Sky Solar Racings. During Summer 2020, I took APS360 and designed an AI Project classifies ingreient photos and recommends recipes based on ingredients and dietary restrictions.
Besides engineering, I am a creative writer since I was 6. You can find some of my works on Medium I am also passionate about sketching, dancing, and music.
I am passionate about software, embedded systems and robotics.
Feel free to get in touch with me if you are interested in my work.
projects
I am trying to explore my field by doing projects in teams. If you have any questions about my projects, feel free to contact me.
Feel the Grid
A pattern memory game for visually impaired people
"Feel the Grid" is a pattern memory game for visually impaired people. Think about a memory game where you try to remember a pattern you just saw,
and you have to recreate the pattern. Instead of trying a pattern you "saw", we use auditory signals to indicate where a "colored" tile
is in a 3*3 grid as you hover your hand through the grid. Afterwards, you try to replicate the same pattern by placing your hand over which
coordinate you remember to be colored. As you develop through the game, the pattern gets harder.
Our game was written in verilog, and uses the following:
- LFSR: We used line feedback shift register(LFSR) to select from different patterns we have for each difficulty level stored in memory.
- VGA: Even though the game is tailored for visually impaired people, we still wanted audiences to be able to follow the game. VGA displays the 3*3 grid and updates as the user receives a pattern or inputs their tiles.
- Visual Sensors: We used 6 visual sensors, 3 along each axis, and used Arduino to interpret the signals to inform our finite state machine about the coordinates of a hand detected.
- Audio output: We stored three different sounds in memory: hand hovering over colored tile, correct pattern, and wrong pattern. As the user is hovering their hand through the colored tiles, in the pattern recognition stage they hear the first sound. Once they input their pattern and press to see if their pattern was correct, they receive either a correct or wrong pattern sound.
SquadMaps
A Graphical Interface System for you and your squad
SquadMaps is a graphical interface system built in C++ for ECE297: Communications and Design Course. With SquadMaps, you are able to:
- Search points of interests using addresses or intersections, supported with auto suggestions
- Search points of interests by simply double clicking on them
- Find travelling directions and send them to your email address
- Receive weather information through weather API
- Change the UI of the map from light mode to dark mode
- Open websites of points of interests
- Find birdsview distances between several points
- You can find the website of our GIS here
Photographic Recipe Search
Multiclass Classification using AlexNet CNN Architecture Embedded in Web Application
Photographic Recipe Search is a web application which uses AI to identify ingredients from images and recommend recipes to users. My teammates were Daniel Lu, Kooresh Akhbari and Yousif Al-Furaiji
connect
If you would like to get in touch with me, please don't hesitate to reach out.