Staff Engineer, Agentic AI

Wwshemi; San Francisco (CA); Sur place; USD 160 000 - 250 000; Plein temps

Résumé du poste

Wwshemi in San Francisco, CA, seeks a hands-on Tech Lead to drive the development of intelligent agents for desktop engineering workflows. You will shape architecture, connect user research with product priorities, and improve reliability and cost efficiency of real-world tasks.

The role combines leadership and hands-on coding on a small AI team, with a focus on tool calling, context management, and robust evaluations to ensure agent success in engineering software environments.

Qualifications

• 7+ years software engineering with hands-on AI experience.
• Proven ability to ship LLM-based agents that perform real-world actions.
• Strong Python skills and familiarity with tooling for evaluation.

Responsabilités

• Own task success metrics and improve completion rates for agent workflows.
• Build evaluation infrastructure based on user stories and workflows.
• Set budgets and track costs for completed tasks and workflows.
• Collaborate with researchers and engineers to map and validate workflows.
• Translate stories into testable evaluations and prioritize coverage by value.
• Make architecture decisions for tool calling, state management, and routing.
• Provide technical direction, review designs, and contribute production code.
• Align agent behavior with product needs and customer workflows.

Connaissances

LLM architecture

Python

Code review

Leadership

Evaluation tooling

Outils

LLM Tooling

Python tooling

Observability

Tracing

Tool APIs

Description du poste

About the Role

Lead the development of agent intelligence that helps mechanical engineers complete complex, multi-step workflows across desktop engineering software. As a hands-on technical lead on a small AI engineering team, you will shape the agent architecture, connect user research with product development, and improve the reliability and cost efficiency of real-world workflows.

What You'll Do

• Own agent task success metrics, establish performance baselines, and systematically improve completion rates.
• Build rigorous, reproducible evaluation infrastructure grounded in validated user stories and real engineering workflows.
• Set per-task token budgets and track the cost of completed workflows.
• Work with researchers and engineering domain experts to map, interview about, and validate user workflows.
• Translate validated user stories into testable evaluations and prioritize workflow coverage by customer value and technical feasibility.
• Make architecture decisions across tool calling, state management, error recovery, model routing, and context management.
• Set technical direction for a small team, review designs and code, unblock teammates, and contribute production code.
• Collaborate with product, integrations, and customers to align agent behavior with real-world use.

What We're Looking For

• At least 7 years of software engineering experience, including 2 or more years building and shipping LLM-based agents that take real-world actions.
• Deep experience with LLM application architecture, including model selection, context management, retrieval, tool calling, and orchestration.
• Strong Python skills and familiarity with function calling, tool APIs, tracing or observability, and evaluation tooling.
• Experience building benchmarks for task completion, cost efficiency, and failure analysis.
• Hands-on technical leadership and code review experience on a small engineering team.
• Experience applying AI or LLM tooling to proprietary engineering data or desktop engineering software, with background in mechanical engineering, CAD, CAE, PLM, or a related domain.
• Familiarity with desktop automation and enterprise workstation constraints is valuable.

Compensation & Benefits

Salary range: $160,000 to $250,000 USD annually. Visa sponsorship is not available.

Location

On-site in San Francisco, California, United States.