Curious to know more about PPO? “People at PPO” is a blog series that highlights the work our team does every day. In each post, we interview one of our team members to get an inside view of PPO and to learn about their experience working here. Join us as we tell their stories and discover how their work impacts the technology we build.
In this post, we chatted with Kesha Bodawala, a Software Developer at PPO. We asked her about her experience with machine learning and working from home. Here’s what she had to say:
J: Hi Kesha, thanks for taking the time to chat about your experience!
Kesha: No problem! It’s nice to meet you as well.
J: Let’s start with what you do at PPO?
Kesha: I work as a Software Developer. My main job is to train machine learning models. I train models with the data from the spectrometers and RGB cameras to assess the quality of products like chicken. Aside from that, I also work on automating the Smart Imaging System, Red Green Blue (RGB) cameras and Programmable Logic Controller (PLC). This means I coordinate the cameras and spectrometers with the PLC to make sure that the system rejects the product. I’m also interested in working on the web app which I’ve been helping to build.
J: Building the web app sounds cool! I’m curious, what are you currently working on with the web app, and what makes you interested in working on it?
Kesha: I mostly work on the front end and how the web app looks. I’m really interested in how we can use it to improve our models. The web app is used to look at what’s happening in the factory. We used our System to find foreign objects. Sometimes they turn out to be false positives. We see that information on the web app and change our models. What I find most interesting is how changing our models provides a better customer experience.
J: That’s so cool! What do you think the future of the web app will look like in one to two years?
Kesha: Right now, there’s a lot of development work with the web app. We’re going into the web app every day to look at its findings. In the future, we’re building it to automatically tell the customer the findings in a way that’s easy to read. It would show the customer the total amount of false positives and foreign objects that it found that day. We’re also working on a way to make recommendations on what the customer can do differently based on the data the System finds.
J: That’s super exciting! I know the customers would love to have everything automated. What does a typical day for you look like?
Kesha: Because of the pandemic, I’ve been working from home. I typically start my day at 9 am going through my emails. Overnight there’s some machine learning processing that happens. After I check my email, I spend most of my morning looking at the processing results from the AI and planning the next steps. Then, I’ll spend most of my afternoon logging and analyzing results in detail and see what is working and what is not!
J: Oh that’s how you train machine learning! Could you give an example of a machine learning project you work on?
Kesha: Yeah, so one project we worked on is foreign material detection. We detect each pixel in the spectral image as either background, product like chicken or pork, or as foreign material. If the model detects enough pixels as foreign material, we reject. For those models, we focus on two rates: true positives and false positives. True-positive is how many times we can detect a foreign material correctly, and false-positive is how many times it rejects false clean product.
J: It’s interesting to hear how you train the AI on false positives. Before you joined PPO, could you talk about your experience with machine learning?
Kesha: I did my undergrad in electronics engineering, but my first internship was doing image processing and computer vision algorithms. That was my first introduction to machine learning. I became very interested in it because it was different from what I was learning in school at the time. It wasn’t classical computer vision or math algorithms, which I enjoyed. I then went to the University of Waterloo for my master’s to study machine learning. At the time, machine learning was very different from what I’m doing now. But I learned how to use different libraries and languages in the program. That has been the most important skill I use at PPO.
J: You also mentioned earlier that you work from home, I was wondering if you could talk about your experience with that?
Kesha: Like most people, when the pandemic hit, the first few days were difficult. We had to set up the VPN from home and learn how to access it. We also struggled with communication. Sometimes two people on the software team would be solving the same problem which became an issue. We learned we had to over-communicate who was working on what. So we implemented more stand-ups and created a new software page on Confluence where we could see who’s working on what. Since working out all the kinks, working from home has been a better experience.
J: Yeah, I definitely relate to the hard transition from working in the office to working from home. I’m glad to hear that you’ve learned to work through the transition. Moving on, what’s your favorite thing about working at PPO?
Kesha: There’s a lot of things I like about PPO. The first is the people. I live alone, so it’s really nice that Sierra and Laurie check in with me. The software team also has rotational chats. During those chats, we talk about our social life and whatever we’re working on. That’s been very helpful for me. The other thing I like about PPO is our virtual events. My favorites were the virtual escape room and cooking sessions. Before our virtual dumpling-making event, I moved into a new apartment with a bigger kitchen. I was so happy I got to use the new kitchen during that event.
J: That’s really exciting that you got to use a new kitchen! I also really enjoyed the dumpling-making event. My favorite part is the banter we had with the team while we made the dumplings. One last question, what’s your favorite food?
Kesha: Chicken wings. When I joined PPO, I was really excited to learn that I got to work with chicken wings. I also thought that since I was looking at them for days and weeks, I would get tired of them. So far, that hasn’t worked, and I still love chicken wings. My favorite place to get chicken wings is Morty’s. Now, I like to cook my own chicken wings at home. I bake them and add a bit of Frank’s Red Hot Sauce.
I also like to pair them with a Belgian Moon. This is because when my parents first came to Canada, I took them to McCabe’s in Uptown Waterloo. At McCabe’s, I ordered chicken wings with a Belgian Moon. When I pair the wings with a Belgian Moon, I have this emotional connection to that moment.
J: Aw I love that story with the Belgian Moon. It’s really sweet to have an emotional connection with your favorite food. Anyways, thanks again Kesha for sharing your story and work experience at PPO.
Kesha: You’re welcome! It was great chatting with you.
J: You as well!
Interested in working at PPO? Check out our Careers Page for more information!