Agentic AI Developer – Python

Photon; Boston (MA); On-site; USD 150,000 - 210,000; Full time

Job summary

Photon in Boston seeks an Agentic AI Developer – Python to design and deploy autonomous AI agent systems using large language models and multi-agent orchestration. You will build and integrate agents with enterprise tools, create planning and reasoning workflows, and deploy on cloud platforms.

Qualified candidates will have 5+ years of Python, 2+ years in LLM-based development, and hands-on experience with FastAPI, Pydantic, asyncio, and multiple AI frameworks.

Qualifications

• 5+ years of hands-on Python development experience.
• 2+ years of experience with LLM-based application development (prompt engineering, chains, agents).
• Strong proficiency in Python frameworks: FastAPI, Pydantic, asyncio.
• Experience with at least one agentic AI framework: PydanticAI, LangGraph, LangChain, CrewAI, AutoGen, or Semantic Kernel.
• Working knowledge of LLM APIs: OpenAI, AWS Bedrock, Azure OpenAI, Anthropic.
• Experience building RAG systems with vector databases (Pinecone, Weaviate, FAISS, OpenSearch).
• Familiarity with MCP (Model Context Protocol) for tool/service connectivity.
• Understanding of agent patterns: ReAct, function calling, tool use, planning/decomposition, multi‑agent orchestration.

Responsibilities

• Design and develop AI agents using Python frameworks such as PydanticAI, LangGraph, LangChain, CrewAI, AutoGen, or Semantic Kernel.
• Build multi-agent workflows with planning, reasoning, reflection, and tool-use capabilities.
• Integrate agents with enterprise tools via MCP, REST APIs, and SDKs.
• Develop and deploy agentic applications on AWS Bedrock AgentCore and/or Azure OpenAI.
• Implement RAG pipelines for grounding agent responses.
• Design structured outputs using Pydantic models for validation and schema enforcement.
• Build evaluation frameworks (LLM-as-Judge, unit testing, compiler validation) for agent quality.
• Implement observability, tracing, and monitoring for agent workflows in production.
• Collaborate with platform engineering, security, and DevOps teams for production deployment.
• Apply responsible AI principles, guardrails, and security controls to agent systems.

Skills

Python programming

LLM-based development

Asyncio

API design

Tools

FastAPI

Pydantic

LangChain

PydanticAI

LangGraph

CrewAI

AutoGen

Semantic Kernel

OpenAI API

AWS Bedrock

Azure OpenAI

Anthropic

MCP

Job description

Job Title: Agentic AI Developer – Python

About the Role

We are looking for a skilled Python developer with deep expertise in building autonomous AI agent systems. You will design, develop, and deploy agentic AI solutions that leverage large language models (LLMs) to autonomously plan, reason, and execute complex tasks using tool integrations and multi-agent orchestration.

Key Responsibilities

• Design and develop AI agents using Python frameworks such as PydanticAI, LangGraph, LangChain, CrewAI, or AutoGen
• Build multi-agent workflows with planning, reasoning, reflection, and tool-use capabilities
• Integrate agents with enterprise tools via Model Context Protocol (MCP), REST APIs, and SDKs
• Develop and deploy agentic applications on AWS Bedrock AgentCore and/or Azure OpenAI
• Implement RAG (Retrieval-Augmented Generation) pipelines for grounding agent responses
• Design structured outputs using Pydanticmodels for validation and schema enforcement
• Build evaluation frameworks (LLM-as-Judge, unit testing, compiler validation) for agent quality
• Implement observability, tracing, and monitoring for agent workflows in production
• Collaborate with platform engineering, security, and DevOps teams for production deployment
• Apply responsible AI principles, guardrails, and security controls to agent systems

Required Qualifications

• 5+ years of hands-on Python development experience
• 2+ years experience with LLM-based application development (prompt engineering, chains, agents)
• Strong proficiency in Python frameworks: FastAPI, Pydantic, asyncio
• Experience with at least one agentic AI framework: PydanticAI, LangGraph, LangChain, CrewAI, AutoGen, or Semantic Kernel
• Working knowledge of LLM APIs: OpenAI, AWS Bedrock, Azure OpenAI, Anthropic
• Experience building RAG systems with vector databases (Pinecone, Weaviate, FAISS, OpenSearch)
• Familiarity with MCP (Model Context Protocol) for tool/service connectivity
• Understanding of agent patterns: ReAct, function calling, tool use, planning/decomposition, multi‑agent orchestration