AI Engineer | Global Commodities Trading | Private AI & Multi-Agent Systems - Cititec Talent Ltd

Build AI that puts complex data to work in global commodities trading.

We’re partnering with an established international energy and commodities trading firm to hire an AI Engineer. This is an opportunity to develop private AI capabilities within a business where data, analytics and the ability to act on information are central to its competitive edge.

The appeal is the scope. You’ll work across the full engineering lifecycle: selecting and adapting open-source models, improving inference performance, connecting models to enterprise knowledge, and building agents that can execute complex workflows.

For an engineer who enjoys taking an idea through experimentation and into production, this offers substantial technical depth and a clear business purpose.

The opportunity

You’ll help turn advances in LLMs and agentic AI into secure, reliable tools for the business. Working with engineering, data, infrastructure and business teams, you’ll identify useful applications and build the systems needed to deliver them within private infrastructure.

Your work will span:

1. Private model development: Evaluate, fine-tune and deploy open-source LLMs, adapting them to specialist datasets and business requirements.
2. Agent engineering: Build agents that use tools, retain context, plan tasks and coordinate across multi-step workflows.
3. Enterprise knowledge: Develop retrieval systems that connect models to internal documents and data through embeddings, vector search, reranking and effective context management.
4. Performance engineering: Improve GPU utilisation, memory efficiency, throughput and response times through quantisation, batching and optimised model serving.
5. Production reliability: Establish evaluation, monitoring, access controls, guardrails and human oversight so systems can be used confidently.
6. Business integration: Connect AI capabilities to internal applications, databases and APIs, taking solutions from initial experimentation through deployment and ongoing improvement.

What you’ll bring

You’ll combine strong Python and software engineering skills with practical experience building and deploying LLM applications. You should be comfortable making decisions across model quality, infrastructure performance and production reliability.

We’re particularly interested in experience with:

1. Hosting and adapting open-source models such as Llama, Mistral, Qwen, Gemma or DeepSeek.
2. Fine-tuning techniques including SFT, LoRA, QLoRA or other PEFT approaches, supported by rigorous evaluation.
3. PyTorch, Hugging Face and GPU environments, including CUDA and inference tools such as vLLM, TensorRT-LLM or llama.cpp.
4. RAG pipelines, vector databases, semantic retrieval and reranking.
5. Tool-calling agents, state and memory management, and orchestration using frameworks such as LangGraph, LlamaIndex, AutoGen or custom implementations.
6. Building maintainable services with APIs, containers, automated delivery and production monitoring.

Why consider it?

This role brings together three compelling engineering challenges: making open-source models useful for a specialist business, making them perform efficiently on private infrastructure, and enabling agents to act reliably across enterprise systems.

You’ll have the opportunity to contribute across model development, retrieval, orchestration and deployment, while building expertise in the commercial environment of global energy and commodities markets.

If you want to turn deep technical work into AI capabilities people can use in their day-to-day work, we’d like to hear from you.

Apply to discuss the role and business in confidence.