AI Engineer- Dallas, TX

Photon Career Site

About The Position

We are seeking a highly skilled AI Engineer proficient in Python to lead the technical development of our Agentic AI platform. In this role, you will move beyond simple prompt engineering to build sophisticated autonomous systems. You will be responsible for designing the architecture that allows agents to plan multi-step tasks, access external databases via RAG, and interact with third-party software through function calling. The ideal candidate treats LLMs as a component within a larger, robust software system, prioritizing reliability, scalability, and observability.

Requirements

Expert Python Skills: Deep experience in asynchronous Python, Pydantic for data validation, and FastAPI for building robust service layers.
AI Frameworks: Hands-on experience with LangChain, LlamaIndex, or specialized orchestration libraries.
LLM Expertise: Deep understanding of LLM capabilities (OpenAI, Anthropic, Gemini) and local model deployment (Ollama, vLLM).
Data Infrastructure: Proficiency with vector databases (e.g., Pinecone, Weaviate, Milvus, or Chroma) and traditional relational databases (PostgreSQL).
Engineering Best Practices: Experience with CI/CD, Docker, and monitoring tools to ensure AI agents are production-ready, not just "demo-ready."

Nice To Haves

Experience with multi-agent systems where different agents have specialized roles and hand-off protocols.
Contributions to open-source AI projects or a strong portfolio of Agentic AI experiments on GitHub.
Knowledge of fine-tuning techniques (LoRA, QLoRA) for specific domain tasks.

Responsibilities

Agent Orchestration: Build and maintain complex agentic workflows using frameworks like LangGraph, CrewAI, or AutoGen.
Tool & Skill Integration: Develop Python-based tools and "plugins" that agents can invoke to perform real-world actions (e.g., querying SQL databases, interacting with APIs, or executing code).
Advanced RAG Pipelines: Architect and optimize Retrieval-Augmented Generation (RAG) systems using vector databases to provide agents with long-term memory and domain-specific knowledge.
Reasoning & Planning Logic: Implement and fine-tune reasoning patterns such as React (Reason + Act), Chain-of-Thought, and Plan-and-Solve to improve agent reliability.
System Evaluation (Evals): Build automated testing frameworks to measure agent performance, accuracy, and "drift" using tools like LangSmith or custom evaluation harnesses.
Performance Optimization: Optimize for latency and cost by managing token usage, implementing intelligent caching, and selecting the right model for the right task.