Head of Engineering, AI Platform

About SAIL

Starburst helps the world's largest enterprises access and analyze their data wherever it lives. SAIL is the team behind AIDA, Starburst's AI product that lets enterprise users ask questions about their data in plain English and get trusted answers. Our AI runs in some of the most demanding enterprise environments in the world, which means it has to meet a bar most AI products never face. Answers must be correct, safe, auditable, and fast.

About the role

SAIL Platform in India starts with you. This is not a satellite team. You will own a charter, hire the engineers who deliver it, and be accountable for outcomes that customers feel.

The charter is the agent platform, and it has three arcs.

1. Internal agent platform. The foundation Starburst's own teams build agents on today, spanning the agent SDK, evaluation frameworks, guardrails, and model serving infrastructure. Your users are Starburst's own engineers, and their feedback lands in hours, not quarters.
2. Extending it toward customers. Take what proves itself internally and harden it for customer-facing use cases. That means multi-tenancy, security boundaries, and APIs stable enough to run in environments we do not control.
3. Observability. What an agent did, why it did it, what it cost, and how well it performed, for both audiences. Internal teams need it to debug and improve; external customers need it to trust agents with production workloads.

You serve two customer bases in sequence, internal agent teams first, external enterprises as the platform matures.

You will be a player-coach. In the first months you will build alongside the first hires you make. As the team grows, you will shift toward leading, but you will never be far from the code.

As the Head of Engineering, AI Platform at Starburst you will: 

• Own the agent platform roadmap, from internal foundation to customer-facing capabilities, and deliver it
• Hire and grow a team of senior engineers in India, and set the technical bar through your own work
• Make architecture decisions for the agent SDK, evaluation, guardrails, serving, and observability, with the internal-to-external evolution in mind from day one
• Run the platform like a product in production, with an on-call rotation, incident response, and a culture of root-cause analysis over quick patches
• Operate autonomously across time zones, making the call when counterparts are offline, documenting the reasoning, and aligning when they come back online
• Work with US-based product, agent, and research teams as peers who shape decisions together, and with engineering leaders across Starburst to redefine how we build, test, and ship software with agents in the loop
• Engage directly with enterprise customers when platform capabilities are on the line

Some of the things we look for:

• You have led an engineering team through a full cycle, hiring it and raising the bar with each hire, shipping with it, and growing the people on it. Typically this means 10+ years of engineering with 3+ leading, but we weight the evidence over the years
• You have built platform or infrastructure products that served multiple engineering teams, and you can talk concretely about the interface contracts, the adoption curve, and what you intentionally left out
• You have run LLM systems in production and know the difference between a demo and a system enterprises trust. That includes the unglamorous parts of LLMOps, timeout and retry design, token cost discipline, latency budgets, capacity, and the different ways models fail
• You can articulate what you adopted, what you waited on, what you built yourself, and why, in a field that changes every quarter
• You are product curious. You ask who the user is and what they are trying to do before you ask what the architecture should be, and you have sat with users of what you built and changed course based on what you saw

Nice to Have:

• Experience taking an internal platform external, turning something built for your own engineers into a product customers pay for
• Experience with agentic systems, evaluation methodology, or LLM serving at scale
• Comfort across Python and JVM ecosystems
• Experience with data systems, SQL engines, or analytics products
• Experience running multiple models behind one platform (proprietary APIs and open-weight, routing and gateways, per-model quirks), including deployments in regulated environments
• Active presence in the ML/AI community in India (meetups, conferences, open source, writing), which helps us find great engineers and helps them find us