Software Development Engineer II: 430

Amagi is an AI-enabled industry cloud platform built for the new video economy, helping media companies modernize operations, unify streaming and broadcast workflows, and drive advanced monetization. From live remote production and real-time ad decisioning to automated playout and global content syndication. We operate a dynamic content and advertising marketplace and offer fully managed broadcast services that ensure 24/7 reliability and compliance. Trusted by 56% of the Top 50 media companies globally, Amagi powers over 7,000 channel deliveries across 300+ content distributors, processing 500K+ hours of content and generating 26 billion+ monetized ad impressions. Headquartered in Bengaluru, India, with a 1000+ member global team across Americas, EMEA and APAC, Amagi is redefining how media is created, distributed, and monetized—intelligently and globally. **Our identity and mission:** Amagi, epitomizing the essence of freedom, embarks on an extraordinary mission to establish the world's foremost media technology business rooted in a foundation of goodness. Our commitment to autonomy is paralleled by our shared connection through a compelling purpose, with the Amagi Way serving as our guiding light. In our pursuit, we strive to create a harmonious blend of individual freedom and collective purpose, shaping a unique and transformative journey that sets us apart in the realm of media technology. Work Mode                   : In Office Location                        : Bengaluru (Bannerghatta Road) **Job Profile:** This is a deep backend engineering role on the infrastructure that runs autonomous AI agents in production. The problems here are the genuinely hard kind: durable, long-running workflows that have to survive restarts, failures, and hours-long executions without losing state or acting twice; correctness that has to hold across distributed service boundaries; concurrency and shared state under failure; non-deterministic model behavior wrapped in deterministic guarantees; and multi-tenant isolation that cannot be allowed to leak. You'll own modules end to end — not just add endpoints, but chase root causes through several services and data layers to find them, and ship into a system where "mostly works" isn't good enough. It's a large, dense codebase with real engineering depth, and you're expected to get up to speed fast and operate without hand-holding. If shallow CRUD work bores you and you want problems with actual weight to them, this is that. **#What you'll do** - \*\*Debug hard, cross-cutting problems\*\* — the kind that don't reproduce in one service, where the cause is three hops upstream in a workflow, a race, or a state-serialization bug, and the fix has to be correct under concurrency and failure. - \*\*Design and implement features across the backend\*\* — from public API through business logic to persistence — that hold up in a durable, multi-tenant, production system. - \*\*Take full ownership of modules that need dedicated engineering bandwidth\*\*, becoming a genuine second owner of parts of the system rather than a ticket-taker. - \*\*Harden what already exists\*\* — close correctness gaps, improve failure behavior, and raise the reliability bar on paths that carry real production traffic. - Write clean, tested, maintainable code — including the automated tests that make a system this complex safe to change. - Work directly with the platform's technical lead on direction, and with operational and application teams on what to build and why. **## What we're looking for** - 2–4 years of professional software engineering experience with a \*\*strong developer mindset\*\* — you reach for root causes, you can own a problem end to end, and you ramp fast on code you didn't write. - Solid \*\*Java\*\* (Java 17+ ideally) and comfort with a JVM backend stack. - Working knowledge of \*\*REST and/or gRPC APIs\*\*, \*\*relational databases (PostgreSQL)\*\*, and writing correct, testable service code. - Comfort operating in a distributed, multi-service architecture — you understand that a request crosses boundaries and you can reason about failure across them. - The ability to work effectively without heavy supervision, and the judgment to know when to ask. **## Our stack (so you know what you're walking into)** Java 21 Dropwizard 4 Vert.x \*\*Temporal\*\* (durable workflows) gRPC PostgreSQL + JDBI Redis Kafka Weaviate Kubernetes + Helm Maven. Notably \*\*not\*\* Spring — if you've only ever worked in Spring Boot, you'll be fine, but expect to learn our patterns.