AI Internship Hackathon

OneInbox AI Internship Hackathon Build. Ship. Get Hired. Think you have what it takes to build production-ready AI systems? The OneInbox AI Internship Hackathon is a hiring-first engineering challenge where exceptional builders compete for paid internship opportunities across AI Engineering, Voice AI Engineering, and Backend Development. This isn't a hackathon for slide decks or AI wrappers. You'll solve real engineering problems inspired by production systems and showcase your skills directly to the OneInbox engineering team. Internship Opportunities We're hiring for: AI Engineer Intern Voice AI Engineer Intern Backend Developer Intern 💰 Stipend: ₹20,000–₹25,000/month (based on performance) Hackathon Format Round 1 — Online Architecture Challenge Design your solution before writing code. Choose a problem statement and submit: Product Requirements Document (PRD) Full Stack Architecture Diagram Solution Overview No implementation is required in Round 1. 📅 Timeline Registrations Open - 2nd July Round 1 Starts - 4th July | 9:00 AM IST Submission Deadline - 14th July | 9:00 PM IST Results - 15nd July | 9:00 PM IST Round 2 — Offline Hiring Hackathon 📍 OJONE, Bengaluru 📅 Timeline Round 2 Starts: 15th July | 10:00 PM IST Development Phase: 15th July – 24th July Final Submission & Presentation: 25th July | OJONE, Bengaluru Results: 25th July 📅 25th July 🕙 10:00 AM – 8:00 PM Shortlisted participants will build a complete production-ready AI agent using their Round 1 architecture as the foundation. The agent should be completed by 24th July. On 25th July, participants can make final enhancements before presenting their AI agent live to the OneInbox Engineering Team and judges. Final internship selections will be made after the offline evaluations. Problem Statements Participants can choose one Problem Statement from the following domains: AI Engineer Problem Statement 1 - Self-Correcting RAG Pipeline : Build a RAG system that identifies insufficient or conflicting context and intelligently re-queries or asks for clarification instead of hallucinating. Measure improvements by comparing hallucination rates before and after introducing the self-correction layer. Problem Statement 2 - Constrained Structured Extraction at Scale : Extract structured JSON from noisy real-world documents while strictly following a predefined schema. The solution should achieve high schema accuracy at scale without relying on hardcoded regex-based extraction. Voice AI Engineer Problem Statement 3 - Interruption-Aware Voice Agent : Develop a real-time voice agent that gracefully handles user interruptions, updates context, and continues conversations naturally. Demonstrate seamless multi-turn interactions with interruption handling in action. Problem Statement 4 - Low-Latency Multilingual Voice Pipeline : Build a multilingual voice-to-voice pipeline capable of switching languages mid-conversation without restarting the workflow. Optimize the system to deliver fast, low-latency responses with real performance measurements. Backend Developer Problem Statement 5 - Rate-Limited Multi-Tenant API Gateway : Design a scalable API gateway that supports multiple AI providers with tenant-specific rate limiting, failover, and intelligent request queueing. Validate the system through concurrent load testing and reliability benchmarks. Problem Statement 6 - Event-Driven Pipeline with Guaranteed Delivery : Build a fault-tolerant asynchronous job processing pipeline with retries, idempotency, dead-letter queues, and observability. Ensure jobs are never lost or duplicated, even during worker failures. Open Innovation Track Problem Statement 7 - Have your own idea? Build an original solution under AI Engineering, Voice AI Engineering, or Backend Development that matches the technical complexity of the official challenges. Submission Requirements Round 1 Product Requirements Document (PRD) System Architecture Diagram Solution Overview Round 2 GitHub Repository Working Applica