Applied AI Engineer, Agentic Systems

ShipBob, Inc.

10h•Remote

About The Position

As an Applied AI Engineer at ShipBob, you will lead the shift from manual workflows to autonomous operations. You won’t just be building chatbots; you will be architecting agentic systems that reason, use tools, and execute complex business logic independently. This is an "AI-First" engineering role where you will embed directly with business teams to map out operational friction and rapidly deploy end-to-end agentic solutions. From designing multi-agent orchestrations to building the "glue" code that connects LLMs to our core logistics APIs, you will be responsible for operationalizing AI at scale. We are looking for a high-agency builder who thrives in ambiguity and is obsessed with the transition from "software that waits for input" to "agents that take action."

Requirements

You think like a founder, thrive in ambiguity, and are excited to challenge assumptions to deliver impactful solutions.
Agentic Expertise: Deep understanding of modern agent architectures (e.g., ReAct, Plan-and-Execute) and orchestration frameworks (e.g., LangGraph, CrewAI, or Semantic Kernel).
Engineering Rigor: 1–3 years of software engineering experience (or a standout portfolio of agentic projects) with strong proficiency in Python and .NET.
LLM Proficiency: Hands-on experience with LLM APIs (OpenAI, Anthropic) and a proven ability to translate raw model outputs into structured, deterministic actions.
Systems Thinking: Experience building retrieval pipelines (RAG), working with vector databases, and designing clean APIs that agents can easily consume.
Full-Stack Mindset: Ability to build the lightweight dashboards and interfaces (Vue.js/React) necessary for humans to interact with and monitor your agents.
Bias for Action: You prefer a working prototype over a long slide deck. You are comfortable challenging assumptions to deliver impactful, automated outcomes.

Responsibilities

Architect Agentic Workflows: Design and implement autonomous agents capable of handling multi-step reasoning, tool-calling, and error recovery.
Operational Discovery: Embed with logistics and support teams to identify high-impact opportunities for agentic intervention.
End-to-End Ownership: Own the entire agent lifecycle: Discovery → Agent Architecture → Prompt Engineering/Fine-tuning → Hardening → Production Deployment.
Build for Reliability: Develop robust guardrails, evaluation frameworks, and logging systems to ensure agents perform predictably in a high-stakes supply chain environment.
Contribute to the "Agent Hub": Document and publish reusable agent modules and toolsets to accelerate AI enablement across the entire organization.
Rapid Iteration: Bridge the gap between "cool demo" and "production tool" by shipping small, gathering feedback from domain experts, and iterating fast.