AI Engineer (mid or senior)
[Position Code: AIE202611]

Job Overview

Date Posted

November 10, 2026

Location

Remote

Experience

3-5+ Years Experience

Qualification

Degree in related field, or equivalent practical experience

Job Skills
Python (core)
FastAPI / Sanic
various LLMs
SQL
LangGraph
LangChain or similar
Prompt Engineering
MCP servers
REST APIs & API Gateways
auth
Text-to-SQL
PII detection & anonymisation
RAG
Prompt-injection defense
Evaluation
Docker
Send your CV to [email protected]

Job Description

Superbo builds AI solutions for businesses. We design, build and run AI agents that connect safely to our clients' systems and data, on our own platform and on custom projects.

We ‘re looking for an AI Engineer, mid-level or senior, to design, build and run LLM-based agents in production. You'll split your time between delivering projects and improving the platform that delivery runs on, with delivery first. You'll own parts of a project from design to production, on your own or alongside the team, and make sure what you build is reliable, measurable and safe with personal data.

You'll join our AI team and work alongside engineers, data specialists and project managers to deliver agent-based solutions. Some of that work extends our own agent platform; some is built from scratch for a client's needs. We pick the tools and models that fit each project.

This is a remote position.

What you will do:

  • Design and build LLM agents: prompts, tools, memory and orchestration, with LangGraph, LangChain or similar frameworks.
  • Choose the model that fits each task and project, across providers and open models, on quality, cost and latency.
  • Integrate agents with external systems through APIs, API gateways and MCP servers.
  • Build retrieval (RAG) over documents, PDFs and structured data, and improve answer quality.
  • Measure what agents do: build evaluation sets, track quality, cost and latency, and fix regressions.
  • Protect personal and client data in everything sent to models, for example with anonymisation (Presidio) and access controls.
  • Scope and estimate parts of a project with the team, and deliver them end to end.
  • Improve our platform: design reusable components, track new models and tools, and run small experiments to decide what we adopt.
  • Write production-grade, tested code, and review your colleagues' code.
  • Deploy and monitor agents in production (containers, cloud, logging), and help handle incidents.
  • Senior candidates: mentor colleagues and help shape our engineering practices.

What you must have:

  • 3–5+ years of software engineering experience, with 2–3+ years focused on AI/ML or LLM-based systems.
  • You've shipped LLM-powered systems to production, then supported and evaluated them after launch.
  • Python, Git, testing, code review and CI.
  • Fluent written and spoken English.
  • A degree in Computer Science or a related field, or equivalent practical experience.

What sets you apart:

  • Text-to-SQL agents over real relational data, such as product catalogs.
  • LangGraph (or a similar agent framework) in production.
  • Experience across several model providers, including choosing a model for cost, latency and quality.
  • PII detection and data anonymisation.
  • You've designed a system's architecture, not only built parts of one.

Nice to have:

  • MSc in Computer Science, AI/ML or a related field.
  • Vector databases, embeddings, PDF/document search, hybrid search.
  • GDPR in practice.
  • LLM guardrails and prompt-injection defense.

Skills:

  • Languages & frameworks: Python (core), FastAPI / Sanic, SQL
  • Agents: LangGraph, LangChain or similar; prompt engineering, tool calling, tool selection and tool retrieval, structured output
  • Models: Working across providers and open models; choosing a model per task for quality, cost and latency
  • Integration: MCP servers, REST APIs, API gateways, auth (OAuth, API keys)
  • Text-to-SQL: Schema and business-context design, query validation, SQL guardrails
  • Retrieval: RAG, chunking, embeddings, vector stores, hybrid search, re-ranking
  • Privacy: PII detection and anonymisation, data minimisation, GDPR-aware design
  • Safety: LLM guardrails, prompt-injection defense
  • Evaluation: Golden test sets, retrieval metrics (recall@k, MRR), generation metrics (faithfulness, answer relevance), LLM-as-judge, regression evals in CI
  • Observability: Tracing (LangSmith, Langfuse), cost and latency tracking
  • Delivery: Docker, CI/CD, one cloud platform, logging and monitoring

How you work:

  • You turn a need into clear technical tasks.
  • You explain trade-offs to colleagues who aren't engineers.
  • You're comfortable with unclear, changing requirements.
  • You own your work end to end, and flag risks early.
  • You judge new models and tools on evidence, not hype.

If you want to join a team of global leaders in their field, send your CV to [email protected], referencing the position code or click the Apply button above.

All applications will be treated in strict confidentiality.

Only suitable candidates will be contacted.