AI Engineer

This is a fully remote position, open to applicants in India.

📋 Description

• Take ownership of the agent loop, which encompasses step control, planning, retries, error feedback, subagents, and parallel tool calls.

• Develop narrow, well-defined, permissioned, and secure tools, MCP servers, and skills for agents working with customer data.

• Create agent context utilizing compaction, memory, retrieval-as-a-tool, and strategies for processing extensive enterprise records.

• Integrate golden sets, regression suites, and trace-based scoring into CI to assess every prompt, model, or harness modification.

• Direct routing across models based on capabilities, latency, and cost; assess new model launches and recalibrate the harness.

• Establish runtime capabilities, including durable sessions, sandboxed execution, nested-trace observability, token budgets, and graceful fallbacks.

• Design backend services comprehensively, covering data models, APIs, storage solutions, and failure scenarios.

• Construct asynchronous pipelines that incorporate queues, workers, retries, idempotency, and backpressure.

• Enhance systems for scalability through caching, batching, indexing, and managing p95/p99 latency budgets on multi-tenant AWS.

• Implement tenant isolation, least-privilege access protocols, PII management, and audit trails.

• Produce design documentation, advocate for technical trade-offs during reviews, and take ownership of services in production.

• Manage a Studio backend service, deliver production updates, and participate in its on-call rotation.

• Analyze agent failures using traces and deploy fixes that significantly enhance task success rates.

• Launch new agent capabilities or backend services for active customer use with established target metrics.

⛳️ Requirements

• A minimum of 3 years of experience in building production backend systems with a strong focus on system design.

• Proven experience in designing and scaling services under real-world load conditions.

• Familiarity with selecting data stores, modeling data, and managing asynchronous processing, queues, caching, idempotency, and production failure scenarios.

• Proficiency in Python is preferred.

• Experience in shipping at least one LLM agent to real users.

• Ability to articulate an agent's loop, tools, failure modes, and performance metrics.

• Developed evaluations that influenced product or engineering decisions.

• Strong instincts in context and tool design; capable of diagnosing agent failures through trace analysis.

• Utilizes coding agents such as Claude Code, Codex, or Cursor as a fundamental part of the workflow.

• Demonstrated history of owning outcomes, selecting metrics, enhancing them, and providing honest reporting of results.

• Nice to have: experience with agent harnesses, frameworks, or orchestration layers; enterprise B2B experience in regulated domains; experience as a founder or founding engineer; familiarity with LangGraph, Agent SDKs, or Deep Agents.

🏝️ Benefits

• Competitive salary along with bonuses and equity options.

• 49 days of paid time off.

• Comprehensive benefits package.

• Retirement and wellness plans.

• Support for professional development.

• Bonus for remote workspace setup.

• Annual company offsites in exciting locations around the globe.