Requirements
- Extensive experience (5-10 years) in backend architecture and design for scalable, distributed systems using Python and modern frameworks (FastAPI, Django REST Framework, Flask).
- Proficiency in building asynchronous code. Must have good knowledge of a web framework like FastAPI, DRF, or Flask, with specific, hands-on experience using asyncio to build scalable, I/O-bound services.
- Proven expertise in architectural decision-making, evaluating trade-offs, and designing robust, maintainable solutions for complex business problems.
- Strong background in code review, enforcing best practices, mentoring team members, and driving continuous improvement in code quality and maintainability.
- Deep experience in multi-agent system workflow and design, including orchestration, communication protocols, and agent lifecycle management using frameworks such as LangChain, AutoGen, CrewAI, or similar.
- Advanced skills in database schema design and optimization, including relational (PostgreSQL, MySQL) and NoSQL databases, with a focus on scalability, normalization, and performance.
- Expertise in API design, including RESTful, asynchronous, and event-driven APIs, ensuring security, scalability, and maintainability.
- Experience in breaking down complex requirements into actionable tasks, creating detailed work breakdown structures, and aligning deliverables with business objectives.
- Strong skills in effort estimation, resource planning, and risk assessment, ensuring timely and predictable delivery of high-quality solutions.
- Strong proficiency with Python testing frameworks like pytest, with a focus on writing comprehensive unit, functional, and integration tests.
- Solid understanding of Python packaging, dependency management, and virtual environments, with hands-on experience using tools like Poetry, uv, pip, and virtualenv/venv.
- Strong understanding of basics of SQLreading and writing SQL queries; a basic understanding of database interaction tools, schema design, and database optimization.
- Hands-on experience with Python data libraries (Pandas, NumPy).
- Good knowledge of API development and testingincluding but not limited to HTTP, RESTful services, Postman, and allied cloud-based services like API Gateway.
- Should have a keen eye for architecture. Understand the trade-off between architectural choices, both on a theoretical level and an applied level.
- Good exposure to LLM SDKs (e. g., OpenAI, Anthropic, Azure OpenAI, Google Gemini).
- Understanding of LLM orchestration and lifecycle management, including prompt engineering, agent state management, and debugging agentic loops.
- Familiarity with Retrieval-Augmented Generation (RAG) patterns and practical experience with vector databases (e. g., Pinecone, Weaviate, ChromaDB, or pgvector) for managing long-term memory and knowledge bases for agents.
- Strong grasp of Agentic AI concepts, including the ability to design, build, and orchestrate autonomous agents that can reason, plan, and execute tasks using a predefined set of tools.
- Experience with multi-agent systems and frameworks (LangChain, AutoGen, Google ADK, and CrewAI) or building complex chains and agentic workflows.
- Familiarity with emerging open standards for AI interoperability, including the Model Context Protocol (MCP) for secure agent-tool communication and the Agent2Agent (A2A) protocol for multi-agent collaboration.
- Strong understanding of at least one cloud platform (AWS, GCP, or Azure) to deploy, manage, and scale applications.
- Strong proficiency with Git for version control, including hands-on experience with collaborative workflows on platforms like GitHub or Bitbucket (e. g., branching, pull/merge requests, and code reviews).
- Experience in presenting Proof of Concepts (POC) findings, including performance benchmarks, potential risks, and strategic recommendations to both technical and non-technical stakeholders. Proven ability to translate successful POCs into wellarchitected, scalable, and production-ready solutions.
- Good to have hands-on experience with AI coding assistants like GitHub Copilot and familiarity with agent development platforms such as Google's Agentspace or similar tools.
This job was posted by Nikitaseles Pinto from Hashedin by Deloitte.
