One Year of Traversing the “e ˣ” Graph
Hey! So you must be wondering, “What the hell is e ˣ doing in the title? It must be a scientific article!” Well, to put your mind at ease, this is not an article about the hidden beauty of e ˣ (although that would be an excellent one :P). Instead, it is about a now sophomore’s year of life at IITR. Please allow me to take you to a flashback journey through time!
PS: If you just want to know about my internship experience, skip to the end of the article :)
The start at -∞
I still remember when I first set foot inside IITR on July 20, 2019, full of enthusiasm, excited to be finally at an IIT after multiple years of hard work. I got myself officially registered at IITR at the MAC (don’t get your hopes too high, it isn’t named after the “Apple’s MAC,” it just means Multi-Activity Centre 😛), and then came the step to set up my hostel room (the part I was least excited about 😆).
Then a piece of news struck me as if it were lightening. I was to stay in a double room!!! 😲. That was unexpected to happen with a CS student, and my reaction was:
Although, as I look at it now, it wasn’t a terrible thing to live in a double room 🙈.
I won’t talk about the mess food or about hardships about living in a hostel for the first time. You must be so relieved, right? Let’s skip to July 24, 2019: the day classes commenced. So like every other studious child, in the beginning, I was focused to even study hard at college 🙈 . Still, then as time progressed I realized like every other college student, college life is more than just academic work, it is about making connections with seniors, meeting new people, trying out different things, not worrying too much about any responsibility (don’t take the pure literal meaning though :P) and living as you wish to.
The acceleration for monotonically improving
So how did a non-social guy who knew nothing about computers, programming, real-world problem solving, or any programming language adjusted to a CS guy’s life?
As it turned out, several choices helped me a lot in getting accustomed to the environment at IITR. The most potent of those was contacting seniors and asking them for advice without hesitation. Although the way I chose might not sit well with many people, it did work out for me 🙈. Before you start running your evil minds high 😛, I contacted them through FB and didn’t do anything harmful 😆.
As expected, not everyone had a welcoming reply. Still, in the process, I made a valuable contact that played a huge, huge role in me pursuing the Deep Learning field. One day I got a reply from Aarush Gupta, now a CS graduate. I had contacted him, seeing various research internship experiences on his profile. I knew he was one of the “Machau” people at IITR 😎. I was excited to talk to him and got to learn a lot about IITR, how to excel at CS (not taking too much load), and most important for my interest in AI: how to begin the journey with Deep Learning!
Another boost was a more formal one. All the facchas (or newcomers) were allotted mentors by a body named SMP (Student Mentorship Program) at IITR. I was the mentee of Natansh Mathur, another “Machau” and a research enthusiast in Quantum Computing. He told us a lot about the many helpful resources and various fields in CS, which helped me explore quite a lot of and make informed decisions.
Traveling up the slope
Alright, now I had the resources and advice to start my journey. So the first field I chose to explore was Quantum Computing due to my interest in physics and quantum mechanics. I knew just the right person to go for guidance (yes! you know that too :P), studied that for three weeks along with the classes, and the intense maths got me there, and sadly I haven’t tried it again but will do it soon!
The Geek Gazette SAGA
Around the same time, the magazine groups’ recruitments at IITR were going on, and I decided to try my luck as a designer for Geek Gazette, the tech magazine group at IITR. I didn’t know anything about design. Still, anyways I attempted as there was nothing to lose :). Then GG was the place where I transformed from being a non-social guy to an enough-social guy. The culture at GG helped me a lot in coming out of the shell and making many more friends and connections. The fun we had at chapos (or parties), was something I experienced for the first time. I owe my ability to dance (very poorly) at parties to GG 🙈. Other than this, I discovered that I had an interest in design as well and made bonds with people I hope to continue to remain stable for a long time.
Alright, so you must be thinking: “JJ is a DL enthusiast, where is DL in this article?” Stay with me, soon the two meet :P
The mid-sem exams passed by, and we knew that the strategy to start studying ten days before exams, works well. Now came the time when various Tech groups at IITR started organizing various intro talks and workshops. The period between mid-September and November was the one when I tried a ton of things: Web Dev, Android Dev, CTFs, Game Dev, etc. to name a few. Of these, the two which caught my interest were Deep Learning and Android Dev.
The Vision and Language Group SAGA
So how did I start with my DL journey? It all started with attending the basic discussions organized by VLG. This open group aims to foster the DL research culture at IITR. I started attending the discussions casually, in the beginning, not reading all the material provided by them. Still, after a or two discussions, it caught my interest. Then I started attending those discussions prepared for the topic reading the introductory blogs and research papers (didn’t understand a word :P). Till mid-October, I was sure about diving deep into DL, so contacted Aarush (YAY! he is back!) again asking for resources to begin my journey with DL. He helped me a lot with the resources.
Thanks to the resources, I was able to pass the interview and became a core member at VLG (in January).
The Mobile Development Group SAGA
Now, at the same time, I was fascinated by another field: Android Development. Thanks to various workshops organized by MDG, I decided I wanted to try it out as well. MDG organized an event by the title: “Season of Code” based on the concept of GSoC. We were asked to submit a proposal for our app and then code the whole winter bringing it to life. That is what happened. The result was Societyfy: An app made for finding company at any time for anything. Download it from here.
It also meant me becoming a member of MDG as an App Developer (in January).
By the end of January, I was a part of 3 campus groups with entirely different skills of interest (Design, App Development, and Deep Learning Research). Now, this may sound a bit hectic and overwhelming to manage, along with academics. Fortunately, I was able to manage well among the activities of all the groups. It taught me one more lesson: How to manage time to participate in various college activities efficiently.
It is not all about work
Alright! So far, I have talked about my work, interview, etc. experience. Now it is time for the fun talks! It is not an overstatement that people at college have the most fun, along with doing other things.
I will keep it short cause it is always better to experience than listen to it :P
So, there are ample opportunities and ways to enjoy your life at IITR. The best path is “Chapo.” What does it stand for? Initially, it started as an acronym for chai-pakoda. Now it means everything from a treat from seniors, friends, and big parties by the campus groups. These are the most fun where we are rescued from the mess food and get to dance, do fun activities, and bond with the seniors well.
Now Roorkee is located near to many tourist spots like Mussoorie, Dehradun, Rishikesh, etc. So one can always have a short trip on any weekend and hang out with friends. I had one of these trips to Mussoorie at the end of the first sem, and it was a great experience.
Apart from the ways mentioned above, there are night outs playing mafia or just roaming around the campus. In short, one gets the perfect environment to enjoy, along with working hard.
Takeaways
So, in brief, the last year at IITR has been an excellent experience for me. I have made some terrific friends, learned a lot about time management, balancing work and fun, and enjoying my time to the fullest.
These are some of the points that made my First year at IITR an experience to remember:
- Have a flexible mind: I remember many people from my batch focused only on coding or CS-related stuff from the beginning of the session. If I had never tried my hand at designing, I would not have been a part of GG, which played a massive role in developing my personality.
- Try out whatever you want to: Also, one tradition or dogma at IITR among the freshmen is to practice coding. A few enjoy it thoroughly, and some do it because there is a lot of hype surrounding it. Had I also stuck to it and not tried my hand at DL or Android Dev, I could have never discovered my interest in the two.
- Never hesitate to ask Seniors: The seniors at IITR are the ones that make it different from other institutes. You can ask them about anything (except CG :P), and they would help you get the right answer or connect you with the right people.
- Make connections with people: Connecting with batchmates and seniors at IITR is an integral part of college life. Always try to find new people and increase your circle.
- Never skip a Chapo: Whatever may be the situation, never skip a chapo, always attend one (if you are invited 😛), and bond with people.
You may have noticed that I have used the terms: connections, people, and seniors a lot. That is because these play a huge role in making your college life awesome at IITR.
When tragedy struck
Alright, everything was going smooth with a few bumps in between until March 8, 2020. It was the time when COVID-19 cases had started increasing in India. It was the Mid-sem break. I had returned to my home, where I was supposed to stay for seven days. It turned out according to the current schedule. I don’t get to return until January 15, 2020. So yeah, a 7-day stay gets converted to 10 months.
Now, as you know, working at home is never the same as doing that at college with one’s peers. So this was a testing time for my career as well as personality. I had to make sure not to get too much carried away having fun with all the free time in my hand and, at the same time, not working too much to find myself disconnected from people.
Still staying on the “e ˣ” graph
During the lockdown time, I had to ensure that I keep growing and don’t waste the opportunity I got to make myself excel in the fields I wanted to pursue. As the authority called off the end term exams, I didn’t have to worry much about the course work and could focus well on DL. I completed various courses provided online on Coursera, and Youtube focused on DL and mathematics. One thing for me to strengthen was the implementation part (I was never entirely into coding 😛). I did some projects, started reading research papers regularly, understanding the maths, and preparing summaries for a few I found inspiring.
Now till the start of June, I had done several projects and made myself comfortable with the theoretical concepts in DL and was looking actively for a project under a Professor. I sent out five emails, I am yet to receive a reply from 3 of them 🙈 , and the other two informed me about no bandwidth. So, I started exploring other fields as well, along with sending out applications to Professors. One day I was looking for internships on AngelList. I saw a posting by FilterPixel for an ML Engineer Intern. I applied, got selected, and finally got my first professional breakthrough in the DL field.
The FilterPixel Experience
So, I started my intern on July 02, 2020, at FilterPixel. To be honest, I was a little afraid of working as a professional. I had no experience, and it is a startup. My first project was to improve the exposure detection model’s performance to classify images as overexposed and underexposed. I was excited to read research papers and achieve our purpose of implementing them. As it turned out, that is what I did, although after cleaning around 15000 images and preparing the dataset on my own for my experiments. I didn’t expect it to take too long to be achieved, but we got stuck in making the model learn how to detect not detect white dresses as overexposure. As it turned out, it took three weeks to get this done so that the customer would be satisfied with the results. All this time, I had been continually running experiments and seeing them fail one after another. But having Aayush Arora and Pratyush Goel there to guide me helped me get me to the solution.
First professional acceleration
As I complete half period of my intern at FilterPixel, there are several things that I learned along the way:
- Handling the pressure and failures: A startup doesn’t expect to wait for one to complete a one week task in three weeks. So at the start, I was quite worried about being stuck in a problem for more than a week. Fortunately, Aayush and Pratyush understood the reason for the delay. They helped me a lot in diagnosing the issues and succeeding eventually.
- Concepts matter more than programming: I was supposed to work with TF and Keras for developing the models. I was anxious initially as I had not worked much with them (I use PyTorch for my purposes). Still, as it turned out, we just need to develop the flow and get the idea cleared out in our minds. Implementing is never a difficult task.
- AI, coupled with humans, is a game-changer: So FilterPixel aims to automate the culling process for photographers to segregate the poor photos from the good ones. It saves a lot of time from their clock for other tasks. Checking thousands of images is a tedious job (I know as I did that for 15000 images 😛).
So all in all, the internship has been a great learning experience for me :)
Conclusion
I had decided not to make this blog too long, but in the end, it became an 11-minute read 🙈. Thanks for your interest in reading the whole of it.
In the end, I would just like to say that the last year has been a great ride for me: transitioning from JEE Coaching to College, living away from home for the first time, from knowing nothing about ML or DL or even CS for that matter to interning in the ML field at the end of a year, from being a non-social guy to attending all the parties, meetings, etc. at IITR has been an excellent opportunity for me.
I can’t wait to return to IITR after this COVID-19 pandemic is over, which I hope is near.
To know more about me check out my website.
