Key Responsibilities
AI Agent Engineering (Core Focus)
- Build AI agents and agentic workflows using LangChain and LangGraph that run in production, not just as demos.
- Design and implement RAG pipelines, including document ingestion, chunking, embeddings,
retrieval and re-ranking over vector databases.
- Build tool-calling agents and MCP servers that let agents act on platform data and APIs safely.
- Implement agent evaluation: test datasets, quality scoring and regression checks so changes to prompts, models or tools can be measured.
- Handle the practical realities of LLMs in production, including latency, cost per request, token limits, rate limits, fallbacks and guardrails.
- Iterate on prompts, context design and agent graphs based on evaluation results and real usage.
- Build backend services and APIs in Python (FastAPI) or Java (Spring Boot) that expose and
orchestrate agent capabilities.
- Integrate agents with existing platform services, databases and event streams.
- Write clean, tested code with unit and integration tests as part of every change.
- Debug issues across agent logic, LLM responses, retrieval quality and backend services.Frontend Engineering
- Build JavaScript/React interfaces for agent features, including chat, streaming responses and result views.
- Integrate frontend and backend cleanly through well-defined API contracts.
- Ship features to production through CI/CD pipelines using Docker, and support them after release.
- Add logging and tracing for agent runs (LangSmith or equivalent) so behaviour can be inspected and improved.
- Work closely with product, senior engineers and data science to turn requirements into working agent features.
- Use Claude Code, Cursor and equivalent AI coding agents in daily work, and review AI-assisted
changes carefully.
THE ENVIRONMENT
Technology Environment
- AI Frameworks: LangChain, LangGraph and Model Context Protocol (MCP).
- LLMs: Claude, OpenAI and open-source models via LLM APIs.
- Vector Databases: pgvector, Qdrant, Pinecone or ChromaDB.
- Evaluation & Tracing: LangSmith or equivalent agent evaluation tooling.
- Backend: Python, FastAPI, Java and Spring Boot.
- Frontend: JavaScript, React and TypeScript.
- Data: PostgreSQL and Redis.
- Infrastructure: AWS and Docker.
- CI/CD: GitHub Actions or equivalent modern CI/CD tooling.
- AI Tooling: Claude Code and Cursor.
WHAT WE'RE LOOKING FOR
Qualifications & Experience
Essential: AI Agents
- Hands-on experience building AI agents or agentic workflows with LangChain and LangGraph,
ideally shipped to real users.
- Practical experience building RAG pipelines with a vector database.
- Understanding of tool calling, prompt and context design, and how agents fail.
- Exposure to evaluating LLM or agent outputs where results are not deterministic.
- 2+ years of software engineering experience building production applications.
- Strong coding ability in Python, with working knowledge of Java and JavaScript.
- Experience building REST APIs with FastAPI or Spring Boot.
- Good fundamentals in data structures, databases and testing.Essential: Ways of Working
- Curiosity and speed in learning new AI tools, models and techniques as the field moves.
- Ability to take a defined problem and deliver working software with guidance on design.
- Clear communication with engineers and product.
- Fluent use of Claude Code, Cursor or equivalent AI development tools in daily engineering work.
- Built an MCP server or MCP-based integration.
- Experience with agent evaluation frameworks or LangSmith.
- React frontend experience, including streaming UIs.
- Fine-tuning or working with open-source and small language models.
- Public projects, GitHub work or write-ups on AI agents.
- IoT, energy, industrial or real-time data experience.
