DataHeist

DataHeist is a four-hour CTF × Machine Learning challenge where teams crack a series of challenges to uncover datasets, identify which data is useful, and build the best predictive model for a real-world problem. Think fast, investigate smart, and turn the right data into the strongest predictions. Event Details: Mode: In-person Venue: Delhi Technological University (DTU) Date & Time: 16th October 2026, 11:00 AM Duration: 4 hours Guidelines: Team Size: 1–3 members. Eligibility: Students from different colleges and specializations are eligible to participate. Multiple teams from the same institution are allowed. Teams may use the internet and publicly available resources for research during the competition. Event Format: Single Round: One continuous, approximately 4-hour competition. Crack the CTF: Solve a set of CTF-style challenges to obtain the datasets needed for the modelling task. The Twist: Not every dataset you uncover will be useful. Some datasets may be intentionally bad, while others contain the information needed to build an effective solution. Build & Predict: Analyze the available data, develop a predictive model, and generate predictions for the given real-world problem. Submissions & Leaderboard: Rolling Submissions: Teams may submit their predictions throughout the 4-hour competition. Evaluation: Each submission will be evaluated against the hidden test data using the competition's specified evaluation metric. Live Rankings: Once evaluated, the resulting score will be used to update the leaderboard, allowing teams to track their performance throughout the competition. Iterate & Improve: Teams can refine their approach and make further submissions during the competition, subject to the competition's submission rules.