Senior GenAI Engineer

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

📋 Description

• Take ownership of agents from start to finish.

• Transition a feature from prototype to production: including orchestrator and sub-agent design, the tools utilized by the agent, system prompts, memory management, and the response contract rendered by the frontend. You will be responsible for writing the code that gets deployed.

• Manage the agent runtime. Design the production environment to ensure it is efficient for the member-facing experience and easy to troubleshoot when issues arise.

• Ensure the tools are reliable. Develop the tool layer that agents rely on, such as search capabilities grounded in our actual catalog, retrieval and reranking functionalities, as well as cart and account actions.

• Prioritize safety. Create a multi-layered safety framework: including input and output guardrails, intent and clarification handling, refusal protocols, and PII boundaries.

• Make quality quantifiable. Advance our evaluation processes: implementing structured checks alongside LLM-as-judge evaluations, and managing a review queue for cases with disagreements among judges.

• Implement instrumentation. Enable agents to be debuggable in production with detailed per-session and per-turn timelines, tool and guardrail traces, as well as visibility into tokens and costs.

• Develop a platform. Extract effective patterns and make them reusable, ensuring that future agents and engineers can inherit runtime conventions, evaluation scaffolding, and default guardrails instead of starting from scratch.

• Enhance team performance. Establish technical direction across agent codebases and infrastructure, while maintaining high standards for design and code reviews.

⛳️ Requirements

• Proven experience in real production environments with LLMs and agents. This is the sole essential requirement.

• Strong judgment regarding critical aspects: grounding and retrieval, tool and context design, memory management, cost and latency considerations, safety protocols, and evaluating the effectiveness of these components.

• Ability to analyze a functioning agent and identify reusable patterns, while knowing when to avoid over-engineering.

• Proficient in Python, along with sufficient knowledge across APIs, cloud technologies, and infrastructure-as-code to take ownership of a feature from model call to deployment.

• A history of establishing technical direction and improving the skills of surrounding engineers.

• Capable of navigating between complex orchestration bugs and providing clear explanations of trade-offs to non-technical stakeholders.

🏝️ Benefits

• 💸 Get paid in USD.

• 🗺 Work remotely: design the life that you want.

• ⛱ Enjoy 15 days of vacation each year from the start date.

• 🎄 16 fully paid Argentinean holidays.

• 🩺 Healthcare Benefit: Monthly stipend to use with your preferred healthcare provider.

• 🗓️ 5-year Sabbatical: After 5 years with CookUnity, you receive a 4-week paid sabbatical.

• 🐣 Paid family leave.

• 🕯 Compassionate Leave: 3-5 days available as needed.

• 🧘🏽‍♀️ Flexible benefits tailored to your preferences: a prepaid card for wellness, learning, food, and more.

• 🤖 AI-forward workplace: enterprise access to ChatGPT and Claude to enhance your productivity and growth.

• 🧑‍🏫 Personalized English coaching.