Vice President

• Aug 2023 - Present

Logistics Team Member

• Sep 2021 - Aug 2023

Hosted a workshop for 40+ middle-and-high-schoolers on the mathematical fundamentals of deep learning at the University of Toronto Mississauga, together building a neural network from scratch with NumPy. Hosted a seminar for 80+ attendees on the fundamentals of computing and electrical engineering at the University of Toronto Mississauga. Spearheaded an AI workshop and hackathon for a group of 35+. Developed Handwritten Digit Classifiers and Stock Predicting RNNs along with students as practical exercises for further learning. Evaluated various AI hackathon presentations on economic feasibility, technical viability, and end impact. Show less

Tech Team Member

• Jul 2020 - Aug 2021

"Our mission at Innovire is to inspire young students to strive and become future leaders and help encourage their passion for STEM." This role involves managing the tech portion of the Innovire Summit, whether that be in the form of creating tutorials, creating exciting projects, or developing kits.

Junior Machine Learning Developer

• Jul 2023 - Present

Executing LLM-based R&D. Executing LLM-based R&D.

Regeneron International Science and Engineering Fair (2023)

• May 2023 - May 2023

2nd Research Award in Cybersecurity ($1500) - United States National Security Agency Research Directorate 4th Award in Robotics and Intelligent Machines ($500) - Regeneron 2nd Research Award in Cybersecurity ($1500) - United States National Security Agency Research Directorate 4th Award in Robotics and Intelligent Machines ($500) - Regeneron

Team Canada-ISEF Selectee.

• Mar 2023 - May 2023

Selected as one of the top 8 individuals in Canada to represent the nation at the 2023 International Science and Engineering Fair. This has been awarded for SPRKD - Saddle Point Recruitment for Knowledge Distillation, a novel knowledge distillation method for deep neural networks shown to outperform current approaches while boosting generalization and decreasing model training time. Selected as one of the top 8 individuals in Canada to represent the nation at the 2023 International Science and Engineering Fair. This has been awarded for SPRKD - Saddle Point Recruitment for Knowledge Distillation, a novel knowledge distillation method for deep neural networks shown to outperform current approaches while boosting generalization and decreasing model training time.

Machine Learning Specialist

• Oct 2021 - May 2023

Developed and deployed ML and DNN algorithms for key client problems (full development lifecycle, from data processing to MLOps).Devised custom data-based scoring algorithms for intelligent Keyword Extraction and automatic training annotation based on word embeddings with Word2Vec and Cosine similarities.Leveraged novel and experimental deep learning methodologies to engineer Keyword Extraction models using Long-Short-Term-Memory-based networks (based on recent scientific literature).Developed a GAN, Gradient Boosted Tree, and MLP based recommendation engine and integrated with Flask APIs, Linux, and AWS (DevOps) for deployment.Recognized for exceptional performance through promotions to roles with greater responsibilities (Intern -> Specialist). Show less

Artificial Intelligence + Data Science Intern

• Jul 2021 - Sep 2021

Actionable.co is a behaviour change platform that uses AI to drastically improve the efficiency of employee-training programs. Here, I'm working with a team of 3 to implement a data science recommendation engine - streamlining this process for those who need it most.

Founder

• Aug 2021 - Jan 2022

My first venture - aiming to skyrocket student productivity through smart, data-based productivity metrics. 🎯The core aim of Crysta was to help decrease student procrastination by tracking their energy levels. Basically, Crysta would record various different metrics throughout the day – things like the number of distracting sites visited, the number of task switches, and the amount of time spent per task – and then, using an algorithm, compile all these metrics into a single number known as an energy score. 🔍Crysta would then look at how these scores changed throughout the day to help determine when you were most likely to be productive, and when you were most likely to be creative. Later on, I even integrated some features that would make it easier to work – such as a tasks list and a pomodoro timer. 💡The original idea for the company came after reading Chris Bailey’s Hyperfocus, a book I’ve talked about in a couple other of my videos. As the author puts it, we each have a unique Biological Prime time – where we’re most productive and energetic. Similarly, we also have a time where we’re least energetic, but more creative. Crysta aimed to be a solution that would measure your energy levels – your biological prime time - and then personalize your daily schedule around these findings Theoretically, this would mean significantly higher productivity for the end user. While the company itself didn't take off, I learned a LOT through the experience. Check out the article, website, GitHub Repo, and Loom below to learn more about the insights and lessons gained from the experience! Show less

Innovator at the Knowledge Society

• Jul 2020 - Aug 2021

"TKS is a global community of the world’s most curious, driven, and ambitious people that want to make an impact. If we can train Olympic level athletes from a young age, why can't we train Olympic level CEOs and innovators?" TKS is a 10-month human accelerator program that cultivates ambition, a 21st-century skill set, and develops foundational mindsets within children aged 13-17 to make a lasting impact on the world. Currently, I'm experimenting the field of machine learning and AI, and how we can use these methods to solve real world problems. Check out my articles on various books and technology at https://medium.com/@aditya.dewan124. Show less