AI Builder at Razorpay… · Bengaluru

Razorpay is one of India’s leading full-stack financial technology companies, powering the way businesses move, manage, and grow money. Founded in 2014 by Harshil Mathur and Shashank Kumar with a simple vision - to simplify payments for Indian businesses - we’ve since grown into a fintech powerhouse driving India’s digital payment revolution. Razorpay powers millions of businesses with a smarter, scalable stack that goes beyond transactions to help them truly build and grow. From building AI-native agentic payments, to AI-assisted fraud detection and real-time risk intelligence to automated reconciliation, smart payouts, and predictive financial insights, we are embedding intelligence across our stack to make money movement faster, safer, and more efficient. In close collaboration with ecosystem partners - including banks, networks, regulators - we are pioneering industry-first solutions that are shaping the next era of fintech Across India, Singapore and Malaysia, our products span everything from seamless checkouts to payroll automation - powering a fintech ecosystem that’s redefining how money moves across Asia. Today, that ecosystem supports everyone from early-stage startups to some of India’s largest enterprises, enabling them to accept, process, and disburse payments at scale while expanding into new ways of managing money more efficiently. Our scale speaks volumes: Razorpay processes $180+ billion in annualized transactions, powering leading businesses like Airbnb, Facebook, WhatsApp, Airtel, CRED, BookmyShow, Zomato, Swiggy, Lenskart, Mirae Asset Capital markets, Indian Oil, National Pension Scheme - and over 100 of India’s unicorns. With strong roots in India and growing operations in Southeast Asia, we are shaping the next chapter of financial technology across the region. We are backed by global investors including GIC, Peak XV Partners (formerly Sequoia Capital India & SEA), Tiger Global, Ribbit Capital, Matrix Partners, MasterCard, and Salesforce Ventures, having raised over $740 million to date. Strategic acquisitions - including Ezetap (POS and offline payments), Curlec (Malaysia expansion), BillMe (digital invoicing), and POP (rewards-first UPI) - along with earlier moves in fraud prevention, payroll, and lending, have further strengthened our platform and widened our footprint across Asia. But what truly sets Razorpay apart is our culture. At Razorpay, ownership is our oxygen - you own what you build, with no micromanagement or red tape, just the runway to make your ideas fly. Learning is a lifestyle - if you’re curious, you’ll feel at home here. People > Pedigree - we hire for attitude, hustle, and hunger more than degrees. Transparency thrives over titles - this is where interns question CXOs and CXOs say “thank you.” Guided by our values of Customer First, Autonomy & Ownership, Agility with Integrity, Transparency, Challenging the status quo and a strong belief that Razorpay grows with Razors,  you’ll be part of a 3000+ strong team building not just products, but the financial infrastructure of the future. We’re hiring someone who does both — and does both at speed. This role sits inside the Central AI Initiatives team within the Marketing Team. You will own the full loop: identify a high-impact business problem, conceive the AI-powered solution, build the production system, ship it, and own the outcome — as a single person. The problems change every few weeks as the company’s priorities shift. One month you’re building a multi-agent system for merchant risk detection. The next, you’re deploying an ops copilot that processes 20,000+ events a day. What stays constant is the expectation that you close the loop end-to-end — from a messy business brief to a live system with metrics moving — in days, not quarters. This is not a research role. It is not a PM role. It is not a data science role. It is a builder role that requires all three — simultaneously. You Know Who You Are You have deep expertise across the full building journey: 1. Decompose business problems from first principles. When a messy brief lands, you break it into data, models, integrations, automations, and human touchpoints — and make the right call on what to build vs. buy vs. defer. 2. Design and build the full system yourself. Architecture, backend services, data pipelines, LLM orchestration, frontend — you own the spec nobody writes for you. You write production code, not demos that die after a meeting. 3. Ship to production and own the outcome. Phased rollouts, alert thresholds, business metrics. When something breaks, you fix it. When numbers are below target, you know Why. 4. Switch altitude — technical depth to C-suite context. Debug a model’s internals in the morning. Present the roadmap to leadership in the afternoon. Review legal terms for an AI partnership before end of day. 5. Multiply AI capability across the organisation. You make teams across Sales, Marketing, Risk, and Engineering better at identifying and executing AI opportunities on their own — through frameworks, reference tools, and workshops that change Behaviour. You’ve been using AI tools for over a year — Claude Code or Cursor is your default environment, not a novelty. You orchestrate agents, build with skills and MCPs. AI assistance is muscle memory. You’re a polyglot — picking up new stacks fast because AI handles the ramp. You thrive in ambiguity and default to action. You will be a good fit if you bring – Founder mindset: you’ve started something (a company, a product, a team) and taken it from zero to live with your own hands. You don’t wait for specs. – 3–5 years building software with at least 1–2 years building AI/ML-powered products in production. Depth matters more than years — the right 2-year person beats the wrong 8-year person every time. – Full-stack capability: backend services, data pipelines (Kafka, Airflow), LLM orchestration (OpenAI, Anthropic, Gemini), frontends (React), and deployment infrastructure (GCP, Docker). You don’t wait for other teams to unblock you. – LLM & agent depth: hands-on with multi-agent systems, RAG, prompt engineering, and production AI pipelines — not just API wrappers. – Product instinct: you ask “why” before “how.” You can scope a business problem down to a shippable prototype and articulate business value to non-technical stakeholders. – Speed & autonomy: you can take a problem from discovery to demo in under two weeks. You move faster than the process and create structure only where it helps. – Influence without authority: you align stakeholders across Sales, Risk, Engineering, and Legal by being clear, credible, and fast — without having direct reports in any of those functions. You will not be a good fit if you You need detailed specs handed to you. You prefer deep specialisation over breadth. You’re uncomfortable with ambiguity. You optimise for code elegance over customer outcomes. You hand off after designing or after shipping. You treat “not my job” as a valid answer. You need a stable roadmap to do your best work. You still thrive living in a 2-week sprint. What the work has looked like Projects change frequently. Recent examples from this team: – Browser-based AI Agent — autonomous web navigation agent that crawls merchant websites, detects payment gateways at checkout, and places test orders. Used by Risk Ops for mystery shopping and fraud detection. Headless browser + LLM-driven navigation + real-time Slack alerting. – Ops Copilot Platform — full-stack employee activity intelligence: Rust-based macOS agents for screen OCR and audio transcription (Whisper), centralised server (Bun + Hono + PostgreSQL), React dashboard with WebSocket streaming, VLM-powered workflow extraction. – Pre-Churn Alerting — Airflow DAGs processing risk events and merchant data to generate daily intelligence alerts delivered to Slack for proactive retention. – Analytics Platform Rollout — end-to-end evaluation, procurement, legal approval, and org-wide rollout of PostHog as the product analytics platform. How to apply We screen on one thing before anything else: evidence that you have built something technically demanding end-to-end and taken it live. – Send a note (300–500 words) on the most technically complex system you’ve built — the problem it solved, the architecture choices, what broke in production and how you fixed it, and what you’d do differently now. – Optionally: a link to something running in production, a GitHub repo with real commits, or a case study. If the note shows the instincts we’re looking for, we’ll reach out within a week. The process is short: one technical problem set, two focused conversations, a final discussion with the team Lead. Why Razorpay 1. Scale that matters — billions of transactions, thousands of merchants. What you build goes live on India’s largest payments platform. You know within weeks whether it Worked. 2. Impact at scale — help transform India’s leading payments platform into the country’s first AI-native payments company. The problems are core: merchant growth, payment routing, revenue expansion — not peripheral. 3. Small team, real autonomy — no bureaucracy between your idea and production. Direct access to the CEO and VP Engineering. If the idea is good, it ships. About Us Razorpay was founded by Shashank Kumar and Harshil Mathur in 2014. Razorpay is building a new-age digital banking hub (Neobank) for businesses in India. Our mission is to enable frictionless banking and payments experiences for businesses of all shapes and sizes. We started as a B2B payments company, and today, we process billions of dollars of payments for lakhs of businesses across India. We are the leading payments provider in the country and have been the first to bring to market most major innovations in the past 7 years. Razorpay currently processes multi-billion dollar payments on an annualised basis, touching 200Mn+ end customers through top clients like Airtel, Goibibo, Yatra, Ola, Swiggy, Zomato, Cred etc. We have raised around USD $740M+ in funding from marquee investors like GIC, Alkeon Capital, Lone Pine Capital, TCV, Sequoia Capital, Tiger Global, Ribbit Capital, Matrix Partners and many stellar angels. Razorpay believes in and follows an equal employment opportunity policy that doesn't discriminate on gender, religion, sexual orientation, colour, nationality, age, etc. We welcome interests and applications from all groups and communities across the globe. Follow us on LinkedIn & Twitter