External events

Rank secured: Global Rank 2. + $3000 Prize

The Jailbreaking Attack Track challenges participants to bypass the safety mechanisms of a guarded LLM using only adversarial prompts. Competitors attempt to elicit harmful, restricted, or policy-violating outputs from a model designed to resist such attacks. The goal is to uncover real-world vulnerabilities in LLM defenses and help drive the development of more robust, trustworthy AI systems.

Participants: Himaanshu, Gowri, Raahul, Kishan and Shivam

Rank secured: Gold Medal

LLMs — You Can’t Please Them All challenges participants to produce short essays (on given topics) that “fool” LLM-based judges. The goal is to find adversarial inputs — essays whose quality is ambiguous or manipulable — that lead different judges (LLMs) to diverging scores.

Participants: Himaanshu, Gowri, Kishan, Akriti, Sangita and Shivam

Rank secured: Rank 40 out of 500 + Poster Presentation in KDD

Meta CRAG-MM Challenge 2025 tasks participants with building multimodal retrieval-augmented generation systems that answer user questions using both images (often from smart-glasses) and external knowledge. The competition evaluates how well models combine visual understanding, information retrieval, and multi-turn reasoning to produce accurate, grounded answers for real-world assistant scenarios.

Participants: Alekhya, Aditi Rai, Diksha, Harshavardhan and Rakshit

Rank secured: Rank 32 (Silver medal)

Jigsaw — Agile Community Rules Classification is a Kaggle competition where participants build a model that predicts whether a Reddit comment violates a specific community rule. The dataset provides community rules, comments, and labels indicating rule violations. The challenge tests a model’s ability to understand nuanced, community-specific norms and support scalable, consistent content moderation.

Participants: Harsh, Darshika, Preeti, Priyanshi and Shivam

Rank secured: Global Rank 5

The Social Deduction Track challenges participants to build AI agents capable of playing hidden-role social deduction games, where success depends on reasoning under uncertainty, detecting deception, coordinating with allies, and persuading opponents. The competition tests whether AI systems can navigate complex human-like social dynamics — including bluffing, trust-building, and strategic communication — to accomplish team objectives in an interactive, multi-agent environment.

Participants: Himaanshu, Kishan, Utkarsh and Subhajit

Rank secured: Global Rank 2

Develop an automated jailbreaking attack to maximize the harmfulness of the LLM outputs for the given prompts.

Participants: Shivam Arora, Kishan Rao Sreenadhuni, Himanshu Devendra Aswal, Jinka Naga Sai Gowri and Raahul Nallasamy

Rank secured: Gold Medal, Global Rank 7

Identify exploits for an LLM-as-a-judge system designed to evaluate the quality of essays.

Participants: Shivam Arora, Kishan Rao Sreenadhuni, Himanshu Devendra Aswal, Jinka Naga Sai Gowri, Akriti Singh and Sangita Bhakat

Rank secured: 17 of 991 teams

This competition uses the key indicator of “N-day retention points” to measure user satisfaction.

Participants: Aakashdeep, Anil Surisetty, Deepak Chaurasiya, Himanshu Chaudhary and Kushagra Agarwal

Rank secured: Silver Medal, Rank 57/3852

Task was to predict short-term volatility for hundreds of stocks across different sectors.

Participants: Sourojit Bhaduri

Rank secured: Silver Medal (2464 teams)

Task was to determine if two products are the same by their images.

Participants: Shreyans Singh

Rank secured: Top 10 teams

Participants: Pranav Poduval, Kushagra Agarwal, Rajesh Kumar Ranjan, Sangam Verma, Karamjit Singh and partnering with AiDA.

Rank secured: 17 of 991 teams

User retention points are a challenging problem. To improve users’ personalized product experience and better allow users to enjoy customized entertainment services, this competition uses the key indicator of “N-day retention points” to measure user satisfaction.

Participants: Aakashdeep, Anil Surisetty, Deepak Chaurasiya, Himanshu Chaudhary and Kushagra Agarwal​​​​​​​

Rank secured: 6 of 450+ teams

The task was to forecast daily COVID-19 spread in regions around the world.

Participants: Abhishek Garg, Lalasa Dheekollu, Sonali Syngal, Diksha Srivastava and Yatin Katyal

Rank secured: Semi-finalists in the top 100

The task was to predict server failure.

Participants: Karamjit Singh, Kamal Kant and Deepak Yadav