Introduction (Personal / Team story)
This is the story of Shubhlabh Shrivastava, an ambitious tech enthusiast pursuing a B.Tech degree, specialising in Artificial Intelligence from Medi-Caps College in Indore. He identifies as an engineer by choice, motivated by the need to connect theoretical learning with real-world applications by building projects independently. His passion for computers started in childhood, leading him to be known as a "computer insect" (computer ka keeda).
His recent specific AI use case was sparked by a job posting from Viraj Sheth (CEO of Monk Entertainment, parent company of BeerBiceps). The post sought people to spend six hours daily tracking trends as "doom scrollers" on YouTube and Instagram. Shubhlabh recognised this repetitive task offered "earning, but no learning" and decided this laborious human job could be automated efficiently by AI out of sheer curiosity.
Problem Overview
Q: What is the problem you are trying to solve, or what innovation are you working on?
The problem I am addressing to solve is the reliance on inefficient, repetitive, and growth-limiting human labour for tracking social media trends. This inefficiency was specifically highlighted by the job posting seeking people to spend up to six hours daily as "Doom Scrollers". My innovation, named YouTube Doom, is a simple AI tool designed to serve as a "mood detector for YouTube videos". It automates the identification and analysis of current trends by fetching video titles, descriptions, and comments based on user-provided hashtags and performing sentiment analysis.
Q: Who will benefit from your work?
The primary beneficiaries are organisations and companies that need trend insights but lack the capacity to hire human trend-followers. This includes Movie Production Houses or Marketing Companies that gain significant value by analysing the effectiveness of their paid marketing efforts. I used the example of the movie Saiyara (Siyara) to illustrate how the tool could analyse whether a paid PR campaign generates positive or negative audience sentiment. More broadly, any individual (such as business analysts or researchers) can benefit by using the solution to understand trends in specific domains merely by inputting relevant hashtags.
Motivation
Q: What motivated you to work on this problem or innovation? Why are you so concerned about it?
My's motivation stemmed from a lifelong curiosity about technology and a desire to connect theoretical learning with real-world applications. As an "engineer by choice" and a lifelong "computer insect" (computer ka keeda), the immediate trigger was the job post for "Doom scrollers". He was concerned because this task offered "earning, no learning" and lacked growth potential. He realised that this inefficient human job could be efficiently automated by AI, challenging the conventional reliance on hiring random individuals for repetitive tasks. He built the AI primarily out of curiosity.
Solution Overview - The AI Use Case
Q: Explain the solution in simple words so that anyone can understand.
The solution, YouTube Doom, is simply a tool that functions like a "mood detector for YouTube videos". When a user provides hashtags, the tool studies the video titles, descriptions, and comments, and then performs sentiment analysis to determine whether the overall audience response is predominantly positive, negative, or neutral.
Q: Why did you use AI to solve this problem? Could you not have done it without using AI?
AI was used because it offers efficiency and is productive. I noted that building the project without AI would have been inefficient and unwise. As a fourth-year student, he did not want to waste a year on manual development when AI, utilizing tools like the Replit AI Agent and initial concept validation via ChatGPT, could complete the task in just two or three hours. Using AI allowed him to automate the repetitive job of understanding trends.
Technical Details - How AI is Actually Utilised
Q: Now describe your solution and AI use case in detail, including a technical explanation.
My solution, built primarily using Python, involves three key technical stages:
- Data Collection System: The tool uses the official Google YouTube API to fetch video titles, descriptions, and comments based on user hashtags. This system manages Rate Limiting to ensure YouTube does not block the requests.
- The AI Brain (VADER): The core model is VADER (Valence Aware Dictionary and sentiment Reasoner), a pre-trained model specifically optimised for analyzing social media text. VADER utilises a large Word Dictionary (7,500+ words) and Context Rules to assign a sentiment score ranging from -1 (very negative) to +1 (very positive).
- Analysis and Output: The Doom Index Calculator processes the VADER scores to compute the Basic Doom Index (the percentage of negative posts). The final output, generated using libraries like Plotly, provides visual graphs and Excel/CSV data, allowing production houses to objectively analyse sentiment. The entire setup was hosted on the Replit Platform.
Challenges & Failures
Q: What were the major challenges you faced during this project?
The major challenge was the inability to secure the necessary Instagram API (Application Programming Interface). Since the original "Doom scrolling" job required tracking trends on both YouTube and Instagram, this technical roadblock forced him to limit the project to YouTube analysis only.
Q: Were there any failures in this project?
The key failure was the inability to execute the full vision by integrating Instagram trend analysis. Instead of spending further time on the API issue, Shubhlabh chose to "dismiss" the Instagram feature, meaning the tool did not completely automate the full scope of the trend-following job originally envisioned.
Q: What did you learn from the challenges & failures?
Shubhlabh learned that AI can significantly help people without a strong technical background to efficiently test ideas, as tools like the Replit AI Agent manage "most technical hurdles". The experience reinforced the realization that success requires continuous learning and adaptation, with Shubhlabh warning that if you are "not updating yourself with updated technologies, then soon enough you will get replaced".
AI Readiness Skills
Collaboration Skills
Q: Name the persons or organisations you collaborated with for making this project successful.
Shubhlabh primarily built the project independently, stating he did not collaborate with any person or organisation on the creation itself, noting he "collaborated with Replit AI". However, he leveraged the Replit Platform and Replit AI Agent for technical assistance. His strong collaboration skills are demonstrated by his role as the Technical Head of STIC (Student Technical Club). He also receives guidance from industry seniors, including Sarthak Srivastava (AI Bootstrappers) and Akshay Mandliya (Data Code community), with whom he is currently working on a React project.
Critical Thinking & Creativity
Q: Do you think your solution is the best one for the stated problem?
Shubhlabh critically acknowledged that YouTube Doom is not necessarily the absolute best solution, but it is highly efficient and productive because it quickly validated his core idea. It is optimal for automating the laborious job that offers "earning, no learning" in just two or three hours. However, the best overall solution would require integrating Instagram trend analysis alongside YouTube, a feature he was forced to dismiss due to the API challenge.
Self-Learning
Q: What self-learning did your team do to understand the problem better, or design the solution, or for any other aspect of the project?
Shubhlabh utilized self-learning, validating the initial idea using ChatGPT. He refined the AI approach through online research. He learned technical details from platforms like GFG (GeeksforGeeks) and free YouTube channels, specifically mentioning Code with Harry. This research confirmed that a previously identified AI model was suitable for social media sentiment analysis.
Online Reputation
Q: If you have made any effort to popularise your solution on the Internet, please share the relevant links.
Shubhlabh actively tried to popularize the solution by posting about the project on his LinkedIn and tagging Viraj Sheth. He remains very active on LinkedIn, with nearly 5,000 followers. Additionally, he has applied for a copyright for the project through his college.
Impact So Far & Vision for the Future
Q: What impact has your solution had so far? How do you measure its success?
Though the solution has not yet been publicly posted or deployed, its success is measured by the fact that it successfully validated the core idea. The impact lies in showcasing that a simple AI setup can deliver meaningful insights quickly—data for which production companies typically pay large sums—and that AI can automate the repetitive human job of tracking trends in just two or three hours.
Q: What are your plans for the future of this project? How do you see it evolving?
If the project receives a good response, Shubhlabh plans to deploy it (host it online) and complete the necessary backend work free of cost. He has already applied for a patent through his college. The ultimate evolution involves addressing the previous failure by integrating Instagram trend analysis to fully automate the original "Doom scrolling" job.
Your Advice to Fellow Innovators
Q: What advice would you give to aspiring AI innovators in India?
Shubhlabh advises aspiring innovators, especially non-technical people, to learn the latest AI tools (like replit AI agent, cursor), noting that companies design AI tools with a very good user interface (UI) to remove technical complexity. For coders, he advises first learning a programming language before moving to specialised AI areas. Generally, he emphasizes the critical need for constant learning, warning that if you are "not updating yourself with updated technologies, then soon enough you will get replaced".
Mentor, Support & Inspiration
Q: Do you have a mentor who guided you in this journey?
Shubhlabh attributes his guidance mainly to his seniors. He has benefited from a strong environment of peer support, serving as the Technical Head of his college's first technical club, STIC, for three years. He also stresses the importance of networking, a skill he developed based on senior advice.
Q: Is there a person who really inspires you?
Shubhlabh is inspired by two prominent figures:
- Sarthak Srivastava, Founder of Bit fumes and part of AI Bootstrappers (a community focused on raising tech awareness).
- Akshay Mandliya, founder of the Data Code community, whom Shubhlabh considers his mentor.
So this was the story of young brains who are willing to solve problems by doing. Team of India's Got Intelligence wish them all the best for their future endeavours. Good luck! Keep proving India's Got Intelligence.
