Software Development Engineer III — Backend, Trust & Safety

What we do — Trust & Safety keeps Swiggy safe on both sides: the platform stays safe for our consumers, and the platform and our businesses stay protected from fraudsters and bad actors.

What we build — the real-time risk decisioning platform that sits in the critical path of Swiggy's consumer journeys, detecting and acting on fraud and abuse as it happens, across all of our consumer businesses.

Stack — low-latency Go microservices, gRPC, event streams, NoSQL stores, caching and AWS, with Python and our data platform powering offline analysis and ML-based risk models.

Why it's hard — every decision is a two-sided bet. Act too aggressively and we hurt genuine customers; act too little and the business leaks money. And the call has to be made in milliseconds.

The adversary adapts — patterns that worked last quarter stop working, so the platform has to let risk and business teams respond quickly rather than wait on an engineering cycle.

Your charter — own architectural decisions end-to-end, lead projects independently, mentor engineers, and partner with product, analytics and business teams to translate risk and customer-experience goals into scalable, production-grade solutions.

AI-native by design — we are looking for an engineer who has already put AI systems into production at scale, who builds agentic automations to remove friction for the teams around them, and who can point AI at our own systems at scale to find where they break before an attacker does. If that is how you already work, you will have a lot of room to run here.

Create architectures and designs for new solutions across existing and new areas

Decide technology and tool choices for your team, and be responsible for them

Drive the long-term technology vision for the team

Lead code reviews, design reviews and architecture discussions, and drive engineering best practices

Experiment with new and relevant technologies, driving adoption while measuring yourself on the impact you create

Mentor engineers and raise the overall engineering bar

Design low-latency distributed systems that make real-time risk decisions in the critical path at Swiggy scale

Architect the signal and feature platform — streaming and batch aggregations, entity-level risk features, served within tight latency budgets

Make the platform increasingly self-serve, so risk and business teams can ship new checks and policies without an engineering deploy

Build detection for emerging abuse patterns — entity linkage, velocity and anomaly detection, and ML risk scores wired into live decisioning

Own the business outcome: measure risk exposure against false-positive impact, and build the experimentation and backtesting tooling to prove a change before it goes live

Own reliability in a critical-path service — instrumentation, observability, graceful degradation, and designing for scale from day one

B.Tech / M.Tech in Computer Science or equivalent from a reputed college, with 6 – 8 years of experience in a product development company

Sound knowledge and application of algorithms and data structures, with space and time complexities

Proficiency in Go or Python (ideally both)

Follows industry coding standards, writing maintainable, scalable and efficient code to solve business problems

Strong experience building and operating distributed systems — message queues, event-driven architectures, stream processing, caching, async processing

Solid system design skills, comfortable owning end-to-end architecture

Experience building latency-sensitive services in a transactional critical path — p99s, timeouts, circuit breaking, fallback behavior

Experience with NoSQL databases, Kafka or similar streaming systems, and AWS

Comfort working with data — SQL on a warehouse / data lake, building metrics, and reasoning about decisions from production data rather than intuition

Demonstrated ability to lead projects independently from design to production

Track record of mentoring engineers and elevating team capabilities

What "AI-native" means for this role

Has productionised AI solutions at scale, not just prototypes

Builds automations for real user workflows — spotting where people lose time to manual, repetitive steps and engineering that friction away

Hands-on with open-weight models such as Qwen or GLM to build task-specific SLMs using few-shot prompting

Proven experience in loop engineering and quality gates — designing the agentic loops, evals and automated checks that make AI output trustworthy enough to ship

Can break a system using AI at scale — adversarial testing as a first instinct

Active user of AI coding tools (Claude, Copilot, Cursor) in day-to-day engineering

Worked on trust & safety, fraud, risk or payments systems, with an instinct for how legitimate flows get exploited (ethical hacking exposure is a plus)

Experience with rule engines, policy / decisioning platforms, or configuration-driven systems built for non-engineering users

Hands-on experience productionising ML — feature stores, real-time model serving, model monitoring

Exposure to graph-based detection (entity resolution, linkage analysis) or anomaly detection at scale

https://bytes.swiggy.com/smart-select-tailored-cart-suggestions-38267fdca12b

https://bytes.swiggy.com/automating-mobile-event-verification-1d840f39d300

https://bytes.swiggy.com/improving-video-cache-hits-on-swiggy-apps-610f395fff46

https://bytes.swiggy.com/a-deep-dive-into-dynamic-widget-swiggys-server-driven-ui-system-92cdc3b16ec6

https://bytes.swiggy.com/building-a-robust-mobile-platform-team-2ee40cce8670

Media on Swiggy’s Mobile Apps. Swiggy believes that a seamless and… | by Vignesh Muralidharan | Swiggy Bytes — Tech Blog

https://bytes.swiggy.com/swiggy-design-language-system-1ef9cca11186

https://bytes.swiggy.com/gradle-incremental-test-runner-125cee1e68a7

https://bytes.swiggy.com/swiss-knife-that-powers-the-swiggy-app-dff9dc49a580

We are an equal opportunity employer and all qualified applicants will receive consideration for employment without regard to race, colour, religion, sex, disability status, or any other characteristic protected by the law.