Senior AI Engineer (Agentic AI)

Overview

About the Role:

We are looking for a Senior AI Engineer who can build and ship AI-powered features and workflows using LLMs in production environments.

You will work on integrating AI into real products from APIs and prompt design to agent workflows and retrieval systems. This is a hands-on role focused on execution, iteration, and reliability.

You are not expected to know everything about AI infrastructure, but you should be able to independently build, test, and improve working systems, and guide implementation decisions where needed.

What You’ll Do

• Build and ship AI-powered product features using LLM APIs (OpenAI, open-source models, etc.)
• Design and implement agent workflows (tool usage, chaining, structured outputs)
• Develop and optimize RAG pipelines (retrieval + generation using vector databases)
• Build and maintain backend services / APIs that power AI features
• Improve output quality through prompting, evaluation, and iterative experimentation
• Track and optimize latency, cost, and output quality
• Work closely with Product and Design to convert ideas into production-ready AI features
• Take ownership of end-to-end delivery of AI features (build → test → iterate → ship)
• Provide technical direction on implementation choices (tools, workflows, trade-offs) when required

What We’re Looking For

• 5+ years of experience, including at least 1–2 years of hands-on experience building and shipping AI/LLM-based systems in production
• Strong hands-on experience with: LLM APIs (OpenAI, Anthropic, or similar), Python (preferred) or another backend language
• Solid understanding of: RAG (retrieval + generation) systems, Prompt design, agent workflows, and structured outputs
• Ability to build and ship end-to-end features independently
• Comfortable working in a fast-moving, experimental environment
• Ability to debug, evaluate, and improve AI system outputs in production

What Success Looks Like

• AI features that work reliably in production
• Fast iteration cycles (build → test → improve → ship)
• Measurable improvements in output quality and user experience
• Practical control over latency and cost
• Ability to own features with minimal supervision