Lead Software Engineer, GenAI Products - SVP - Irving

The Opportunity

We're building 'Citi Assist', a Generative AI assistant for every Citi employee on the planet. This is ground‑floor work on a platform that will reach hundreds of thousands of people. You won't be maintaining someone else's system. You'll be shaping ours. This is a hands‑on technical leadership role. You'll split your time between writing code, designing systems, and helping set the direction for how we build.

We want someone who leads by doing, someone whose pull requests set the standard, not just their opinions in meetings.

We sit within Developer Enablement, inside the Chief Technology Office. We own our work from idea to production. No hand‑offs to a separate ops team. No waiting for someone else to deploy. Our culture runs on engineering craftsmanship. Leadership here actually understands what that means. We're proving what's possible inside a large organisation, and that means navigating some legacy alongside building the future.

If you want everything perfect on day one, this isn't the right fit. If you want to help us get there, keep reading. The team is distributed across Pune, London, and Texas. You'll work closely with engineers across time zones and partner directly with product and platform leadership.

• Write code and lead. You're the technical anchor for our backend services. That means contributing high‑quality code regularly, not just reviewing other people’s work, and setting the bar through your own output.
• Own the hard stuff. Core platform capabilities are yours: background jobs, autonomous agents, and the services that keep everything running. You'll build for reliability and performance, and you'll be accountable when things need fixing.
• Shape how we build. You'll have a real voice in our architecture and technical roadmap. You'll work directly with the product owner to make sure we're building the right things, in the right order, for the right reasons.
• Stay ahead of GenAI. This space moves fast. We need someone who follows it seriously, brings new ideas back to the team, and actively applies what they're learning to the products we build.
• Make the team better. Thoughtful code reviews, unblocking people when they're stuck, and sharing what you know. Your success here is tied to the team's success, not just your own output.
• Keep safety non‑negotiable. We're building AI systems inside a global bank. Security, compliance, and AI safety aren’t afterthoughts. You'll lead by example on all of it.
• Communicate like it matters. We're distributed across three continents. Clear, direct, proactive communication isn’t a soft skill here. It’s how the work gets done. You'll model it.

• Proven experience leading technical projects and mentoring engineers.
• Deep, hands‑on experience building and running production‑grade services. You’re still writing code regularly. You’re good at it.
• You’re a polyglot. You’re comfortable across multiple languages and paradigms, whether that’s Python, Java, Go, or Type Script/Node.js, and you pick the right tool for the job rather than defaulting to what you know best. We don’t want someone who only knows one tool.
• Experience designing scalable, distributed systems.
• A solid grasp of software design patterns and a commitment to clean, testable, maintainable code.
• Experience with containerised environments (we use Open Shift, but Kubernetes experience transfers well).
• Practical experience with CI/CD pipelines.
• Solid experience with relational databases, particularly Postgres.

• You lead by doing. Architecture discussions and pull requests—you’re comfortable in both. You don’t just tell people what to build. You show them.
• You build things that last. You make smart trade‑offs between speed and quality. You know the difference between moving fast because it’s right and cutting corners because it’s easy.
• You ship. The goal is delivering value to users, not chasing perfection. You know when to move fast and when to slow down and do it properly.
• You find the best path, not just the obvious one. New code, improved old systems, right third‑party tool—whatever solves the problem. The outcome is what…