Research Engineer

Job Title

Research Engineer

Salary

£70k–£120k + Equity

Company Description

Electric Twin is a London-based AI company that has raised $14M from Atomico and LocalGlobe to build synthetic populations for behavioral prediction. Founded by a former Chief Adviser to the Prime Minister and a military commander, they use LLMs to simulate human responses for global brands with 95% accuracy.

Job Description

You will research and build the cognitive architectures that bring AI agents to life, balancing behavioral science with backend engineering. Working within the Science team, you'll design memory systems and decision-making frameworks to create realistic synthetic populations. This is a high-impact role bridging the gap between LLM research and production-scale infrastructure.

Location

London, UK

Why this role is remarkable

Work on a unique product that uses LLMs to solve the $90B market research problem, achieving 10,000x faster insights than traditional methods.

Backed by heavyweight investors including Atomico, LocalGlobe, and angels like Marc Andreessen and the CTO of Slack.

Direct collaboration with academic leaders from the London School of Economics to publish research and define the future of agentic modeling.

What You Will Do

Design and run systematic experiments to evaluate synthetic agent behavior and validate performance against real-world human data sets.

Build sophisticated cognitive systems using advanced LLM techniques like chain-of-thought prompting, RAG architectures, and multi-agent tool use.

Implement robust memory and personality frameworks that ensure AI agents maintain consistent demographics and behavioral patterns across interactions.

The ideal candidate

Has over 7 years of engineering experience with at least 1 year hands-on with LLMs, production RAG, and agentic workflows.

Possesses a strong research background in quantitative fields like computational social science, physics, or cognitive modeling to simulate human systems.

Proficient in Python and backend frameworks like FastAPI, with the ability to take complex AI prototypes into scalable production environments.