Research Engineer

About Electric Twin

Most organisations make their biggest decisions on a thin slice of evidence — a handful of customer calls, one survey, instinct when the data runs out. Electric Twin changes that. We're a behavioural simulation platform that lets organisations test ideas, messages, and decisions against synthetic populations — digital twins of real audiences — and get answers in minutes that would otherwise take weeks. The work is grounded, and the engineering is real. We've run over 40,000 evaluations across populations covering 155 countries. Independent academic research with Professor Michael Muthukrishna at the LSE found our outputs come back roughly 10,000× faster than traditional methods, at 95% accuracy. AI is part of what we build, but it isn't the product. The hard work is system design, data, orchestration, and making outputs trustworthy enough that people act on them. Running population-scale simulations is a systems problem as much as a science one — foundation models, behavioural data, evaluation pipelines, and the infrastructure that holds them together, with real engineering problems at every layer.

How the work gets done here isn't handed down — it's worked out by the people closest to it. You'd help shape what gets built or sold and why, decide where the trade-offs sit between speed and depth, and judge when something is ready for customers and when it needs another pass. What the role involves day-to-day, what your first six months look like, and how compensation, equity, growth and flexible working are structured are all set out in detail below — the headline is that early joiners get meaningful equity, a clear path to grow with the company, and the flexibility to work around the rest of your life.

The Role

As an AI Engineer you will be part of the wider technical team but positioned within Science. You’ll research and build the systems that bring our AI agents to life and you will work with engineering to scale the infrastructure that powers our LLM-driven synthetic populations. You'll design agent cognitive architectures, implement context engineering and memory systems, while ensuring these AI systems can operate reliably at scale in production environments. This role balances AI agent research with backend engineering. It is ideal for research engineers who want to work directly with large language models to create realistic behavioural simulations, while building the robust infrastructure needed to deploy them in enterprise settings.

What You'll Do

• Research, Modelling and Experimentation: You will design and run systematic experiments to evaluate synthetic agent behaviour, test hypotheses about behavioural patterns, and iterate on model architectures based on empirical results and validation against real-world data.
• LLM Product Engineering: You will build sophisticated prompting strategies, behavioural frameworks, and decision-making systems that enable agents to exhibit realistic human-like behaviour across diverse scenarios and demographics. You will combine this with client interactions to ensure product viability for the end customer.
• Architecture & Development: Design and implement the cognitive systems that give AI agents consistent personalities, memory, and reasoning capabilities, using advanced LLM techniques like chain-of-thought prompting, RAG systems, and agentic tool use.

Who You Are

Essential Qualifications

• Bachelor's or Master's degree in CS, Math, Physics, AI, or related technical field.
• Strong foundation in both AI/ML concepts and backend engineering principles.
• Experience working in fast-paced environments where requirements evolve rapidly.
• Over 7 years of experience with at least 1 year working hands-on with large language models to solve complex problems.

Technical Skills

• Research: Proof of fast iteration and experimentation in order to validate model performance and outputs. Exposure to research-driven product development or academic AI research is desirable. Knowledge of fine-tuning workflows, model optimisation, and experiment tracking. Understanding of statistical validation and data quality assessment. Experience with frameworks for building AI / LLM applications (e.g. PyTorch, Hugging Face Transformers, LangChain).
• LLM & Agent Development: Hands-on experience building applications with large language models, implementing advanced prompting techniques, RAG systems, and agentic workflows. Experience with multi-agent systems, simulation frameworks, or agent-based modelling.
• Backend Engineering: Proficient in Python and backend frameworks (e.g. FastAPI, Django, Flask); understanding of distributed systems and scalable architectures.

Personal Attributes

• Strong ownership mentality—you see projects through from design to deployment.
• Pragmatic problem-solver who balances technical elegance with business needs.
• Clear communicator who can explain complex technical decisions to non-technical stakeholders.
• Thrives in ambiguity and adapts quickly as product requirements evolve.
• Passionate about building infrastructure that enables innovative AI applications.
• Intellectually honest—willing to question prevailing approaches and advocate for better solutions when evidence supports it.
• Collaborative mindset—debates ideas vigorously while respecting other perspectives.

What We Offer

• Competitive salary.
• Meaningful equity in a high-potential seed-stage company.
• Unlimited leave. Take the time you need.
• Generous matched pension contributions.
• Private healthcare.
• Cycle to work scheme.
• Direct access to and collaboration with world-class founders.
• Hybrid working from our London office (4 days in office a week).
• Flexible working around life commitments. We value outcomes over presenteeism.