Devesh Darshan | AIcrew Stories
🎉 Welcome back to AIcrew Stories!
Their stories are sure to inspire beginners in Machine Learning. Read on to get some guidance, learn new tools and tricks. Maybe someday your story will be here! :)
🙋🏼♂️ Let’s meet Devesh Darshan
Meet Devesh Darshan, a third-year student from BITS Pilani. While Devesh is pursuing a Bachelors in Pharmacy course, his interest lies in AI, especially Computer Vision.
🚗 Devesh's Machine Learning Journey
Devesh started his ML journey in the summer vacation after completing his freshmen year. Like every student about to enter their sophomore year, he was exploring different domains. He tried out web development and cybersecurity before getting fascinated by Machine Learning. To better know the field, he took up online courses on Udemy. The course focused more on implementation rather than the theory. He wanted to learn more about what makes these ML algorithms work. To uncover this mystery, he audited the famous Machine Learning course by Andrew Ng on Coursera.
Through his experimentation, he identified his interest in Computer Vision, namely GANs. “I love GANs because the work is not that complex and hard to understand as some other applications in computer vision, but the results produced are just mind-blowing. Initially, all my practice of ML and DL was done on hackathons, later I started doing projects which introduced many new topics to me”, says Devesh.
💪🏼 Getting Started with the Challenge
Devesh found out about ML challenges on AIcrowd through an email from college admins. Detailing his positive experience on AIcrowd, Devesh says, “AIcrowd is one of the most unique DS platforms because of the type of competitions. We can find problems for each domain, be it computer vision, reinforcement learning, audio, NLP or Blitz puzzles”.
Talking about Blitz 8, Devesh said, “The problems in this iteration of Blitz ranged from beginner to advanced level, like from classification all the way to image dehazing”.
🧭 His Approach
Talking about his Community Contribution prize-winning notebook, Devesh had this to say. “I already worked on GANs in Blitz 5 where there was a problem to colour black and white images. I saw the terms like image dehazing and GCANets being used. But I thought why not use a GAN for this problem, so I took the previous code, finetuned some parameters, and it worked”!
📋 Devesh’s Advice for Beginners
Devesh’s advice for beginners in this field is to never stop learning. “Just don’t ever stop learning new things in the field of AI field because each and every day there is some new advancements”, he says.
Sharing some personal experience, he urges enthusiasts to not shy away from math. Adding to this point he says, "Maths might seem boring if you can’t see these algorithms working on real-world data. But once you see these algorithms working and then read the maths behind it, you will find yourself fascinated by how these formulas and numbers can produce these results”. He emphasis the need to practise, "to improve you need to constantly practise, either through hackathons or projects”. “Regular practice, especially in early stages of learning is very important”, he adds.
When not solving college assignments or tackling AI challenges, Devesh likes to game and listen to audiobooks. He enjoys these activities as it helps him relax. He also dabbles in crypto trading.
What stories do you want to read next? Let us know in the comments! Check out more Blitz winner blogs over here.