Razorpay has launched the Razorpay AI Buildathon, a student-only hiring initiative to recruit AI Builder Interns for in-person roles in Bengaluru. The program bypasses traditional resume screening, aptitude tests, and group discussions, shortlisting candidates directly based on practical code, architecture, and working AI prototypes.
Program & Internship Overview
| Attribute | Details |
| Program Name | Razorpay AI Buildathon |
| Role Title | AI Builder Intern |
| Stipend | ₹75,000 / month |
| Duration | 6 Months or 12 Months (Candidate’s Choice) |
| Work Mode | In-Person (Bengaluru) |
| Start Date | September 2026 onwards |
| Selection Mode | Direct technical panel interview based on project submission (No aptitude test / GD) |
| Application Deadline | September 5, 2026 |
Evaluation Criteria
Submissions are evaluated based on four primary parameters:
- Problem Taste: Selecting a meaningful, real-world financial or merchant problem.
- Build Quality: Clean repository structure, execution reliability, and code trust.
- AI Judgment: Utilizing AI models appropriately while opting for deterministic solutions where AI is unnecessary.
- Failure Recovery: Demonstrating resilience by showing what broke during development and how it was resolved.
Submission Tracks & Execution Bars
Track 01: AI Growth & Agentic Commerce
- Focus: Build an agent that grows merchant revenue using Razorpay test-mode APIs, or enables end-to-end transactions for AI buyers.
- Directions: Conversational in-app checkout, agent-readable catalogs, upsell/cross-sell agents, campaign orchestrators.
- Required Bar: Every monetary action must be explainable, bounded, and gated with a clear audit trail and graceful failure handling.
Track 02: AI Risk Manager
- Focus: Develop a detector, verifier, or auto-responder targeting fraud, returns, or chargebacks.
- Directions: Chargeback evidence responder, return-risk scorer, fraud-spike detector, abuse-ring sentinel.
- Required Bar: Strictly defense-only. Must present precise precision/recall metrics on a held-out dataset, including false-positive costs.
Track 03: AI Revenue Recovery
- Focus: Create an agent that detects at-risk revenue (failed payments, cart drop-offs, unpaid invoices) and executes automated recovery workflows.
- Directions: Payment degradation root-cause analyzer, checkout drop-off recovery, B2B receivables chaser, Hinglish voice recovery.
- Required Bar: Demonstrate measurable money recovered across a batch with compliant escalation rules and stopping criteria.
Track 04: AI Finance Controller
- Focus: Build an agent that closes a finance-ops loop across a batch of 50+ synthetic records.
- Directions: Multi-source reconciliation, settlement Q&A agent, forward cash forecaster, tax-line matcher.
- Required Bar: Throughput, verified accuracy, and a documented exception list for unhandled cases.
Track 05: Open Track
- Focus: Solve any meaningful real-world problem utilizing AI in an effective, bounded manner.
- Required Bar: Demonstrate a working product, meaningful application of AI, and clear evidence of value creation.
Application Requirements
Submitting an application requires filling out 12 key details across two sections:
- Personal Information: Full Name, College, Graduation Year, Bangalore Availability (Yes/No), Duration Choice (6 or 12 Months), Resume File.
- Build Details: Selected Track, Project Name, Problem Statement, Public GitHub Repository URL, 5-Minute Pitch Video URL (Unlisted), Explanation of what broke and how it was resolved.