This blog is an effort to serve as a guide for students who are aiming to win Smart India Hackathon. I have decided to pin down my experiences and key takeaways I observed, which I believe were essential in winning this national hackathon. Even though I specify my experiences of SIH’22, some of these might even serve as a guide for most hackathons and students in general, from a fresher/rookie who is exploring university level hackathons to someone in pre-final/final year eyeing a win in hiring/national hackathon.
Following is a step-by-step guide to winning Smart India Hackathon, based on my experience, mentioning key takeaways for each point.
Step 1: Teamwork is the secret to success.
Like other hackathons, Smart India Hackathon requires you to be in a team of 4–6, one teammate mandatorily being a female. Here are some of the things you should keep in mind while forming the team-
- Team composition(based on tech skills)- Each team member should have a different skill set, and their roles clearly defined in the team. In my case, the team comprised 2 members skilled in ML/AI(including me), 2 members who can help in backend development and hosting, and 2 members who can help in frontend, UI/UX, and making attractive presentations.
- Team composition(based on age/experience)- I had the fortune of being in a team led by a senior who had already won the Smart India Hackathon 2020 edition. Although previous SIH winners would most likely not want to participate again, my point of emphasis here is having sincere seniors in your team, can help you a lot, as he/she will maintain the dynamics between having a good team environment for working while also making sure the task doesn’t look too dreadful, by maintaining a light atmosphere. Likewise, if you’re the seniormost teammate, you need to make sure you lead the team gracefully.
- Team composition(based on gender)- SIH usually clearly states that in case of a tie in the final, weightage will be given to the team having more female candidates. While that is not a huge advantage, make sure that the gender balance of the team is good. In my case, it was 3 males, 3 females.
Step 2: The best problems are the ones that are worth solving.
Choosing the right problem statement is another essential checkpoint that you need to clear. While SIH offers many problem statements, some might be more rewarding to you than others.
- Choose what you can solve- The problem statement that you choose should make sure that every member of your team involved can provide valuable input to it and produce the required results. For instance, since both of my teammates knowing ML/AI (me included), were more familiar with Natural Language Processing(NLP) as compared to Computer Vision(CV), we searched for NLP problems only.
- Analyse the difficulty well enough- My team while selecting the problem statements had come to a unanimous decision that problems requiring work in regional languages should be avoided, cause not only they don’t have a good enough corpus available, but also it’ll be hard to find an open sourced Large Language Model(LLM) trained on them if required in the future. So we discarded those problem statements.
- Don’t hesitate in doing the hard work for filtering out problem statements- Even though it might look a bit tedious, I would recommend going through every problem statement once. In my case, I went through all 530+ problem statements, and my team and I filtered out the problems step-by-step before coming to the final problem statement.
Step 3: The first test is a test of your preparation.
Usually, universities keep an internal hackathon round to decide which teams to send for the Smart India Hackathon(in case you are new to hackathons, you should know that you can participate in SIH only on behalf of the university). Clearing this round requires you to usually have a team, a problem statement, a presentation, and some work around it so that the panel is convinced
- No one likes to see code- While having a functionality not integrated with the front-end is good enough for an internal hackathon(depends on competition), ignoring frontend development can cause you to get eliminated at this stage only.
- PPT is the key- While the product you make in the internal hackathon is only for the panel to see if you’re capable enough(you will most probably not even use this as your main product later in the hackathon), your future judgment in SIH, on whether you make it to final or not, depends on your PPT file that you’ll submit on SIH portal once the university selects your team. Make sure that the PPT is concise, contains diagrams, and flowcharts, and is attractive.
Step 4: The second time is the charm.
Remember the purpose of PPT. SIH asks you for PPT to know about how you’re going to approach the problem statement. So, it should be a mix of technical and non-technical stuff. Pasting a screenshot of the frontend you made to show your skills to the internal panel can lead to your elimination at this stage, cause you are only asked to present your ideas, and the work might have been a way to show your skills to the university, but it will be considered as cheating by SIH, as you are allowed to code only during the finale.
Step 5: The final showdown.
Once you get notified that your team has made it to the finale, it’s time to get in action mode. Every mistake at this stage can dent your chances of winning SIH. You need to take care of various stuff before the finale happens and during the finale.
a) Before the finale
- Research stuff about the organisation- The problem statement you selected must be from a domain, an organisation. Research about them, and talk to people who are experienced in this domain. For e.g. my problem statement was about creating a keyword-based search for court cases. So my team consulted a lawyer and asked him for his advice, on how the judiciary, ecourts, etc work and what can we build which helps to solve the real problem. You shouldn’t make the mistake of just trying to make your solution technically attractive. This will help you bring novelty in your solution for the problem statement which will help you stand out from your competitors.
- Explore other problem statements of the same organisation- If you solve your problem statement, while also integrating solutions of similar problem statements in your own, and showing them as a novelty, you increase your chances further.
b) During the finale
There are usually 3 checkpoints, and the judges keep checking your progress, while sometimes casually roaming nearby you and taking interest in your work. Here are a few things you must keep in mind
- Implement suggestions by the judges- If your evaluator took interest in your work, and suggested you some changes, you should note them down and positively make them before the next round. This gives off a good impression to the judge.
- You are being monitored- In many centers, someone will be monitoring your team. If your entire team sleeps the entire night, it gives off a bad impression, as it beats the purpose of the hackathon. While resting is important, make sure you rest turn by turn, other than lunch/dinner.
- Integration is the key- The hardest part of your project will not be making efficient AI models, an efficient backend, or an efficient and attractive frontend. It’ll be putting this all together in one place. You are likely to surpass most of your competitors if all of your functionalities are integrated and working.
- Know how to present- All teammates presenting together is never a good idea. One teammate presents while others chipping in, only when it’s required or a specific technical question is asked by the judges.
Bonus- Never give up
My team was sent to the waiting list initially after the internal hackathon. Later my university was asked to submit more teams so they sent my team in the confirmed list. Thereafter, for the grand finale, my team was again on the waiting list after the evaluation of the PPT by SIH. We weren’t fancying our chances back then, thinking that while other competitors in our problem statement got through, we just edged in. However, we kept working and emerged as winners. My advice is to never undervalue yourself, no matter what the odds are.
That’s all from my side. This was my first blog, if you liked it, kindly give this post applause. Feel free to use the comments section to provide feedback.
Follow me on