Job Title: Toptal: Python Backend Development Talent with RAG and Agentic AI Experience
Source: weworkremotely
Summary
We are seeking an experienced Python Backend Developer to design, build, and deploy scalable AI-powered applications using Retrieval-Augmented Generation (RAG), large language models, and agentic AI frameworks. The role will focus on delivering a production-grade RAG system and AI chatbot that can securely integrate with enterprise data, APIs, databases, and cloud services.
General Information
The organization is developing an AI-powered platform and requires an experienced individual contributor to build its RAG architecture and conversational AI capabilities. The developer will work closely with a distributed team and should be available for several hours of overlap with US working hours.
The project involves designing AI systems that go beyond basic prompt engineering, including multi-step workflows, autonomous agents, vector search, knowledge retrieval, memory management, and tool integration. The solution must be scalable, secure, observable, and suitable for production use.
Technology Environment: Python, FastAPI, large language models, LangGraph, LangChain, vector databases, Azure AI services, AWS, Docker, Kubernetes, and microservice-based architectures.
Tasks and Deliverables
- Design the end-to-end architecture for a scalable RAG system and AI chatbot.
- Develop Python backend services, REST APIs, and microservices using FastAPI or similar frameworks.
- Build document ingestion, chunking, embedding, indexing, retrieval, and reranking pipelines.
- Implement vector search solutions using Azure AI Search or comparable vector databases.
- Design autonomous AI agents capable of planning, reasoning, tool usage, and decision-making.
- Develop multi-step AI workflows using LangGraph, LangChain, CrewAI, AutoGen, or similar orchestration frameworks.
- Integrate commercial and open-source LLMs, including Azure OpenAI, OpenAI, Anthropic Claude, Gemini, and comparable models.
- Implement conversation memory, session management, context management, and agent collaboration patterns.
- Connect AI workflows with APIs, databases, enterprise systems, and external tools.
- Develop asynchronous, high-performance services capable of handling concurrent AI workloads.
- Implement prompt management, structured outputs, guardrails, fallback logic, and model evaluation processes.
- Establish logging, monitoring, tracing, observability, security, and error-handling standards.
- Containerize and deploy AI services using Docker, Kubernetes, Azure, or AWS.
- Translate business requirements into technical designs, delivery milestones, and production-ready AI solutions.
- Collaborate with the wider team while independently owning architecture and implementation decisions.
Required Experience
- 8 or more years of professional Python backend development experience.
- Strong experience designing REST APIs, microservices, asynchronous services, and distributed backend systems.
- Hands-on experience building production-grade RAG applications.
- Strong understanding of embeddings, document chunking, semantic search, vector indexing, retrieval strategies, and reranking.
- Hands-on experience developing AI agents and multi-step LLM workflows.
- Experience with agentic AI frameworks such as LangGraph, LangChain, CrewAI, AutoGen, or comparable platforms.
- Experience integrating LLMs through OpenAI, Azure OpenAI, Anthropic Claude, Gemini, or open-source model APIs.
- Ability to design AI architecture beyond basic prompt engineering.
- Experience integrating AI applications with APIs, databases, data pipelines, and enterprise systems.
- Experience implementing security, monitoring, logging, tracing, and observability for production services.
- Experience deploying containerized applications using Docker and cloud platforms such as Azure or AWS.
- Ability to independently translate business requirements into scalable technical solutions.
- Strong communication and collaboration skills in a distributed working environment.
- Availability for several hours of overlap with US working hours.