Staff Engineer, Agentic AI — Lead AI Workflows (SF On‑Site)

Clera; San Francisco (CA); On-site; USD 160,000 - 250,000; Full time

Benefits offered by this job

Equity

Visa sponsorship not available

Job summary

Clera is seeking a Staff Engineer — Agentic AI to own the core agent intelligence layer that translates engineers’ intent into reliable, cost‑efficient multi‑step workflows across desktop tools. You’ll report to the CTO and lead a small team at the intersection of applied agentic AI, user research, and product delivery.

Based on an on‑site role in San Francisco, you will shape the architecture, drive cost controls, and collaborate with customers during POCs to ensure real‑world utility.

Qualifications

• 7+ years in software engineering with real-world agentic AI experience.
• Deep expertise in designing LLM application architectures and tool calling.
• Strong evaluation instincts for agentic systems and cost efficiency.
• Proven track record shipping AI systems with measurable outcomes.

Responsibilities

• Lead development of the core agent intelligence layer for multi-step workflows.
• Serve as technical lead for a small AI engineering team and collaborators.
• Own product loop from user stories to testable implementations and benchmarks.
• Define evaluation frameworks, baselines, and improve completion metrics.
• Set per-task budgets and track cost per completed workflow.

Skills

Agentic AI

Python

LLM tooling

Team leadership

Architecture design

Software engineering

Tools

LangSmith

Logfire

GAIA

τ‑bench

function calling

Job description

Clera is seeking a Staff Engineer — Agentic AI to own the core agent intelligence layer that translates engineers’ intent into reliable, cost‑efficient multi‑step workflows across desktop tools. You’ll report to the CTO and lead a small team at the intersection of applied agentic AI, user research, and product delivery.

Based on an on‑site role in San Francisco, you will shape the architecture, drive cost controls, and collaborate with customers during POCs to ensure real‑world utility.