Software Engineer - AI Developer Productivity

About the role

Baseten is seeking a Software Engineer focused on AI Developer Productivity to own the internal AI developer platform end to end. In this role, you will architect, build, and operate the foundational layers that make AI-assisted development safe, scalable, and fast for Baseten's product teams. You will design and maintain the agent substrate, including repo-level context, internal MCP servers, and shared skills that encode our workflows into reusable defaults. Success in this position means creating the playbook for an AI-first software development lifecycle where the best path is also the easiest path for every engineer. You will measure the impact of your work through adoption, cycle time, and change failure rate, iterating based on data rather than opinion. This is a platform builder role where your responsibility is infrastructure, not evangelism, and your influence grows by shipping tools teams cannot live without.

Key facts

What you'll do

• Own the internal AI developer platform end to end, including architecture, build, rollout, operation, and measurement.
• Evaluate and integrate third-party AI coding tools such as Claude Code, Cursor, and Codex, while building the context layer that makes them work reliably against our monorepo.
• Build frameworks and scaffolding that let other engineers create their own agents without requiring deep LLM expertise.
• Establish the evaluation practice for AI-assisted development at Baseten, designing eval harnesses that compare configurations against real Baseten tasks instead of relying on subjective vibes.
• Instrument AI tool usage and observe downstream effects on cycle time, review latency, and change failure rate to guide investment priorities.
• Drive adoption through developer experience by shipping good defaults, clear documentation, and low-friction onboarding rather than top-down mandates.
• Embed with product and infrastructure teams to identify where AI genuinely unblocks workflows and generalize those wins into reusable platform capabilities.
• Own the safety layer, including permissions, secrets handling, audit trails, and cost management, ensuring agents operate within strict guardrails.
• Create project templates and onboarding flows that ship with AI tooling pre-configured, enabling new engineers to be productive with agents within their first week.
• Build self-serve infrastructure so teams can author and run their own agents without becoming dependent on a central platform team.
• Implement gateway, authentication, and authorization controls for internal model access, with transparent cost controls and detailed audit logging.
• Instrument and maintain sandbox environments where agents can safely build, test, and iterate without risking production systems.
• Define and maintain CLAUDE.md and AGENTS.md conventions, ensuring architecture and domain context stays accurate as the codebase evolves.
• Provide hands-on support and concrete examples that scale beyond one-on-one conversations, making best practices accessible to the entire engineering organization.

Requirements

• Have 4+ years of relevant industry experience building and enabling AI native SDLC.
• Strong proficiency in Python and/or Go, building tools other engineers depend on daily.
• Hands-on experience with LLMs and agent frameworks, including tool calling, MCP, context management, orchestration, and failure handling.
• You have shipped at least one agentic system that real people used, not merely prototyped or experimented in isolation.
• Deep personal fluency with AI coding tools and well-formed opinions about where they currently break down or misalign with production workflows.
• A platform mindset focused on adoption and self-service, treating internal engineers as customers and preferring a good default over a restrictive policy.
• Solid understanding of developer tooling, CI/CD pipelines, and Kubernetes/Docker fundamentals.
• Comfort with ambiguity, recognizing that best practices in this space invalidate themselves every few months and architectures must evolve accordingly.
• Excellent written communication skills, as a large portion of your leverage will come from documentation, templates, and examples that outlast individual discussions.

Practical notes

• Hours: Full-time (40 hours per week).
• Travel: None required.
• Visa: Sponsorship may be available for qualified candidates.
• Deadlines: Applications will be reviewed on a rolling basis until the role is filled.