Agentic AI Engineer - FDE (Hybrid: NYC/SF/Toronto)

Inventure; San Francisco (CA); Hybrid; USD 185,000 - 240,000; Full time

Job summary

Inventure is seeking a Forward Deployed Engineer to join a Palantir-style operation. You will join a software holding company that scales vertical-market software with agentic AI, embedding into portfolio companies and shipping live, measurable workflows.

You will own end-to-end delivery—from integration and prompts to evals and UI—then tune and hand off once stable, while documenting patterns for rapid repeat deployments.

Qualifications

• Four to six years of engineering experience, with three years at a high-growth startup.
• Prior success at a known startup with lasting impact.
• Experience delivering production-grade LLM or agentic software with evals, autonomy gating, and outcome tracking.
• Daily use of AI development tools and opinions on them (e.g., Cursor, Claude Code, Copilot).
• Operator instincts: read a P&L, work with a GM who may not code, and scope technical problems.
• Based in or willing to relocate to New York, San Francisco, or Toronto; comfortable with 3 days in office per week plus travel; visa transfers available.

Responsibilities

• Drop into a portfolio company, learn its operations, and identify two to three high-leverage workflows for agentic software.
• Build the wedge: integrations, prompts, evals and the UI to land the first agentic loop on the core platform.
• Operate and tune deployed systems after launch, adjust autonomy tiers and hand off when stable.
• Act as the technical face to portfolio-company GMs, setting expectations and ensuring the system becomes their leverage.
• Codify deployment learnings into playbooks to speed up future deployments across verticals.

Skills

TypeScript

Python

React

SQL

Job description

Inventure is seeking a Forward Deployed Engineer to join a Palantir-style operation. You will join a software holding company that scales vertical-market software with agentic AI, embedding into portfolio companies and shipping live, measurable workflows.

You will own end-to-end delivery—from integration and prompts to evals and UI—then tune and hand off once stable, while documenting patterns for rapid repeat deployments.