AI Agentic Engineer

About the Role

We're looking for a multi-disciplinary AI Engineer to design, implement, and deploy LLM-driven agents with strong backend and front-end integration. You'll lead work across five areas: agent harness engineering, agent development, agentic workflow development, LLM inferencing/evaluation/hosting, and LLM fine-tuning — and build real, production-grade applications on top of all of it.

Profitmind builds this stack against real retail data. Our platform runs a team of nine specialized agents (Data Load, Strategy, Competitive Intelligence, Pricing, Inventory, Promotions, Assortment, Planning, and the Monday Morning agent) that analyze a retailer's entire business every week and hand merchandising teams a ranked list of actions with the dollar value attached.

The ideal candidate combines a strong Python and AI foundation with hands-on LLM knowledge (prompt engineering, context management, structured outputs and tool calling, retrieval, evals, and fine-tuning with LoRA or QLoRA), harnesses (Claude Agent SDK, OpenAI Codex SDK, Pi, OpenClaw, Hermes Agent), and practical experience across agent frameworks (LangGraph, PydanticAI, Google ADK).

Key ResponsibilitiesAgent Harness Engineering

• Build and own the agent harness layer, including agent loops, tool execution, context management, session persistence, permission controls, and sandboxed environments.
• Evaluate and extend existing harnesses, including Claude Agent SDK, OpenAI Codex SDK, Pi, OpenClaw, and Hermes Agent, when they offer a better path than building internally.
• Architect agentic orchestration systems that govern complex multi-step planning, long-term memory management, and dynamic tool use.
• Design resilient execution environments equipped with semantic guardrails, fallback mechanisms, and self-correction loops to safely handle API timeouts, context window overflows, or hallucinated tool calls.
• Standardize internal API contracts for tool creation, enabling seamless and secure "plug-and-play" integration of new enterprise data sources into the agent environment.

Agent Development

• Architect and build production-grade LLM agents using frameworks such as LangGraph, PydanticAI, and Google ADK, or custom agent loops when a framework adds unnecessary complexity.
• Develop composable, tool-augmented patterns using retrieval, planning, reflection, and subagent delegation where appropriate.
• Integrate vector databases and knowledge graphs to support retrieval-augmented generation, memory, and grounded decision-making.
• Engineer context deliberately through just-in-time retrieval, progressive disclosure, compaction, structured note-taking, and isolation between agents.
• Develop prompt and tool strategies that are evaluated for reliability and robustness.

Agentic Workflows and Platform Integration

• Build multi-step and multi-agent workflows with routing, parallel execution, checkpointing, retries, compensation, and human approval steps.
• Create durable workflows that can survive restarts, resume safely, and degrade gracefully when dependencies fail.
• Build MCP servers that provide agents with structured, auditable, least-privilege access to internal systems.
• Develop scalable FastAPI services for synchronous, asynchronous, and streaming agent execution.
• Connect agents to internal applications, chat platforms, scheduled and event-driven triggers, and CI pipelines.
• Build agent interfaces using React, TypeScript, Next.js, or similar technologies, including real-time streaming over SSE or WebSockets.
• Design clear user experiences for long-running agents, including progress visibility, interruption, steering, approval, and recovery.

LLM Inference, Evaluation, and Hosting

• Evaluate and integrate managed or self-hosted models based on application requirements.
• Monitor and improve model quality, latency, reliability, and cost using techniques such as caching, batching, and model routing.
• Build task-specific evaluations and regression tests using Profitmind’s real-world retail workflows.
• Instrument agent behavior to capture tool calls, errors, latency, token usage, and cost.

LLM Fine-Tuning and Continuous Improvement

• Use evaluation results to determine whether prompting, context engineering, retrieval, or fine-tuning is the best approach.
• Support targeted fine-tuning experiments using techniques such as LoRA or QLoRA when the expected benefit justifies the effort.
• Help curate and version evaluation and training data derived from representative business use cases.

What You Bring

• Education & Foundation: A bachelor’s or master’s degree in CS (or equivalent) with 2+ years building Agentic applications.
• Production Agent

Experience

1+ year operating LLM agents in production, with hands-on experience using frameworks (LangGraph, PydanticAI) and SDKs/harnesses (Claude Agent SDK, OpenAI Codex SDK).

• Agent Architecture: Deep understanding of agent loops, tool calling, retrieval, and context management, with proven ability to measure and improve agent performance.
• Evaluation & Fine-Tuning: Experience evaluating non-deterministic AI systems and fine-tuning or adapting models for domain-specific tasks.
• Backend & Data Infrastructure: Production experience with FastAPI, Docker, MLOps practices, and vector stores.
• Developer Tools: Proficiency with agentic coding assistants like Claude Code, Codex, or Cursor.
• Education: A bachelor’s or master’s degree in computer science or a related field, or equivalent practical experience.\

What We Offer

• Competitive salary and equity.
• A flexible hybrid working environment.
• The opportunity to shape a production AI platform used to make high-value retail decisions.
• Ownership across the full agent lifecycle, from experimentation and evaluation to deployment and user experience.

Job Type: Full-time

Pay: $100,000.00 - $150,000.00 per year

Benefits

• 401(k)
• Dental insurance
• Health insurance
• Paid time o