Deepfake and online‑fraud challenges go live at IIT Delhi this weekend. Raksham finals run on October 24–25. Thirty shortlisted student teams will build and demo AI tools. They will tackle voice cloning, fake videos, and scripted scam workflows. Recruiters are watching closely for deployable security skills.
IIT Delhi runs Raksham with the Indian Cyber Crime Coordination Centre and Amazon. The hackathon prioritises responsible, real‑world solutions for India’s digital finance stack. The prize pool totals ₹4 lakh across five awards. Standout winners may be showcased at Amazon Sambhav 2026. That visibility can lift resumes and signal job‑ready security skills.
Raksham finals schedule and prize pool
Finale runs for 48 hours on October 24–25 at IIT Delhi. Amazon will sponsor travel for finalists; teams arrange their stay. Teams of two to four are eligible. Prizes split as ₹2,00,000, ₹1,00,000, ₹50,000, ₹25,000, and ₹25,000. Plan logistics early to protect build time.
Raksham challenge themes and judging
Teams choose one track: voice‑cloning fraud detection, scam‑pattern recognition, or deepfake verification. Scoring emphasises problem fit, architecture, feasibility, and responsible AI safeguards. Judges also assess adoption potential and communication quality. Bring clear data lineage, privacy mitigations, and a working prototype plan.
Shortlisted teams and live demo format
Top 30 teams build, test, and demonstrate prototypes before an expert jury. As of October 22, no public finalist list appears on Unstop or club sites. Finalists should monitor registered email and spam folders. Expect a closed demo floor; follow DevClub and ARIES channels for updates.
Why recruiters care: AI‑security skills
Hiring demand is surging for roles securing AI systems and combating deepfakes. Surveys flag shortages in AI‑security engineers and architects across India. Skills that stand out include adversarial testing, robust model pipelines, and privacy‑first design. Strong communication and governance awareness often tip hiring decisions.
How to follow results and prepare for next cohort
Results and clips are expected on DevClub’s X or YouTube, and the Unstop page. Use only consented, public, or synthetic data, and disclose third‑party assets. Practice explainable detectors, escalation paths, and safe deployment choices. Track ARIES and DevClub feeds for the next Raksham‑style cohort.
