What are the responsibilities and job description for the Senior AI Agentic Engineer position at Profitmind?
Profitmind is building the intelligence behind how retailers make pricing and merchandising decisions. Today, many of these decisions are still driven by spreadsheets, rigid rules, and manual judgment, even at the largest brands.
Our platform turns complex data such as sales, inventory, and competitive signals into clear, explainable recommendations merchants can trust. Our platform focuses on impact, helping retailers improve margin, inventory health, and decision quality at scale.
Based in Pittsburgh, Profitmind is backed by a recent strategic investment from Accenture, and scaling its agentic AI platform to power decision-making for some of the world’s largest retailers.
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 be leading efforts across LLM agent design, prompting, fine-tuning, and MLOps, while building real-world, production-grade applications with modern web technologies.
The ideal candidate combines a strong foundation in Python and AI with practical experience in agent frameworks like LangGraph, PydanticAI, and Google ADK, as well as FastAPI and front-end development.
What you’ll do:
LLM Agents & Prompt Engineering
- Architect and implement LLM agents using frameworks like LangGraph, PydanticAI, and Google ADK.
- Build composable, tool-augmented reasoning chains (e.g., RAG, CoT, ReAct, planner-executor).
- Integrate vector databases (e.g., FAISS, Pinecone, pgvector) and knowledge graphs (Neo4j) to support retrieval-augmented generation (RAG) and long-term chatbot memory.
- Design and maintain high-quality prompt strategies for robustness and reliability.
FastAPI, Model Context Protocol (MCP) & Backend
- Develop and maintain scalable APIs using FastAPI, supporting synchronous and asynchronous agent execution.
- Integrate Model Context Protocol (MCP) to enable secure and structured access to external data and tools within agent workflows.
- Implement state tracking, context-aware input dispatch, and modular plugin integration within the control plane.
Evaluation, Testing & Observability
- Build unit and behavioral tests for agents, tools, and workflows.
- Develop tooling for trace analysis, agent state debugging, and hallucination tracking.
- Compare and benchmark agent orchestration frameworks for trade-offs in speed, reliability, and usability.
Model Fine-Tuning & MLOps
- Fine-tune models using LoRA, QLoRA, or full fine-tuning pipelines.
- Integrate, deploy, and monitor models in production using cloud providers.
- Set up agent logging, observability dashboards, and recovery workflows.
Front-End & UX
- Familiar with React, TypeScript, Next.js, or similar frameworks
- Understanding of front-end and back-end integration for AI tools
- Ability to build basic dashboards or agent interfaces
- Integrate agents into interfaces
- Speak the language of UI
What we’re looking for:
- 3 years experience with Python in ML/AI systems and PyTorch or Tensorflow
- 1 years experience with LLM agent development, prompt engineering, and frameworks like
- LangGraph, PydanticAI, and Google ADK.
- Experience with fine-tuning LLMs.
- Familiarity using vector stores like ChromaDB, Weaviate, or pgvector.
- Production experience with FastAPI, Docker, and MLOps
- Expert in Agentic Coding IDEs (Windsurf, Cursor or Claude Code)
- Bachelor’s or Master's degree in computer science
Nice to have:
- Open-source contributions to LLM/agent tooling
- Knowledge of async programming, websockets, and streaming APIs
What we offer:
- Competitive compensation and equity
- Comprehensive benefits including medical, dental, vision, etc.
- Unlimited and flexible PTO