Staff AI Engineer

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

📋 Description

• Create agent systems from foundational principles. Determine the feedback loop, select the appropriate tools, devise the context strategy, and establish the evaluation framework. Make informed choices between single-agent and multi-agent architectures, LLM reasoning versus deterministic post-processing, and retrieval versus direct context loading — and substantiate these decisions with data.

• Engineer the context meticulously. The most challenging aspect of developing an effective agent lies in what is included in the prompt and the resulting outputs. You will focus intensely on context windows, tool interfaces, structured outputs, citation grounding, and the prompt itself.

• Foster rigorous evaluation methods. Construct evaluations prior to developing the agent. Identify failure points, rectify underlying issues, and demonstrate that the adjustments have positively impacted the metrics.

• Utilize AI tools as a power user. A significant portion of your time will be dedicated to utilizing Claude Code, Codex, and similar tools to plan, scaffold, refactor, and troubleshoot your projects. We anticipate you to be more efficient with these tools than most engineers are without them.

• Become a domain specialist. Understanding healthcare claims, coding standards, and medical records is essential in this role. Proficient engineers who immerse themselves in the domain significantly enhance their contributions.

⛳️ Requirements

• 2–4 years of practical ML/AI engineering experience, holding a Bachelor's degree in Computer Science, Mathematics, Engineering, or a related field — or a Master's degree in a similar discipline without prior industry experience required. In either case, you should have developed at least one production-quality system (whether in industry, research, or a significant open-source project) that you managed end-to-end.

• Proficient in Python engineering. You should demonstrate clean abstractions, type discipline, asynchronous programming, and thoroughly tested code.

• In-depth, hands-on knowledge of agent loops — including how a model decides to call a tool, how a tool’s output reintegrates into context, the conditions for loop termination, and common failure points.

• Practical experience with at least one major agent SDK — such as OpenAI Agents SDK, Anthropic SDK/claude-agent-sdk, LangGraph, or an equivalent platform — along with insights into the trade-offs involved.

• Familiarity with the construction of modern coding agents and their context engineering — including what is included in the system prompt, methods for reading and editing files, how long-running tasks are organized and monitored, and where issues may arise.

• Proficient in using Claude Code/Codex as a power user. You should be capable of brainstorming, planning, and executing complex engineering tasks with these tools — including reviewing their source code when necessary to understand or modify behavior.

• Strong command of VS Code and git — including branches, rebases, worktrees, conflict resolution, and pull request workflows. This is essential.

• A strong inclination towards measurement: you do not release without an evaluation, and you do not trust a metric that cannot be reproduced.

🏝️ Benefits

• Work from anywhere in the US! Machinify embraces a digital-first culture.

• Comprehensive Medical/Dental/Vision plans.

• Flexible Spending Account/Health Savings Account.

• Tuition reimbursement program.

• Competitive salary and 401(k) with company matching.

• Unlimited Paid Time Off.

• Additional health and wellness benefits and perks.

• A flexible and trusting environment that empowers you to perform at your best.