Investigate the capabilities, strengths, and limitations of LLMs (e.g., GPT-4, Claude, LLaMA) in specific tasks such as text summarization, question answering, or creative generation.

Analyze open-source LLMs and evaluate their performance against commercial offerings.

Explore domain-specific fine-tuning techniques for LLMs.

Conduct experiments on prompt engineering techniques to improve model responses.

Evaluate the feasibility of integrating visual understanding models alongside LLMs.

Research recent developments and document potential applications in Visualist workflows.

Summarize research findings into actionable recommendations.
‍

Benchmark LLM performance on domain-specific datasets relevant to Visualist’s use cases (e.g., interior design, personal styling).

Test models for specific use cases like generating client feedback drafts or administrative proposals.

Analyze and document how various models handle complex queries, incomplete information, or ambiguous input.
‍

Create prompt libraries optimized for different tasks (e.g., moodboarding, draft generation, client communication).

Develop scripts to test and compare LLM outputs systematically.

Prototype small tools or interfaces that use LLMs to streamline productivity or workflow tasks for Visualist users.

Investigate and test API integrations with commercially available AI models.
‍

Curate, clean, and annotate datasets for model evaluation or fine-tuning.

Analyze user interaction data to identify tasks that could be automated with AI.

Test models on multilingual datasets to evaluate performance in different languages.
‍

You make things happen. Big challenge? Weird problem? You dive in, figure it out, and deliver. You’re not one to wait around—you love moving the needle.

You’re the go-to problem-solver. Friends call you when they’re stuck because you have a knack for untangling messy situations and turning them into clear, actionable steps.

You’re comfortable with ambiguity (and a blank page). Whether it’s a rough idea or no starting point at all, you see it as an opportunity to create something brilliant from scratch.

AI gets you excited. You see its potential to solve real-world problems, especially for creative professionals, and you want to help push the boundaries of what’s possible.

You love experimenting. Whether it’s testing the limits of an LLM or uncovering hidden strengths in NLP tools, you’re all about diving deep and seeing what works (and what doesn’t).

You bring structure to the chaos. Give you a vague or ambiguous task, and you’ll break it down, find the throughline, and make progress.

You care about the details. Testing, documenting, tweaking—it’s not just about getting it done; it’s about getting it right.

You explain complex things like a pro. Analogies? Yes. Breaking big ideas into bite-sized insights? Absolutely. You make it easy for others to “get it.”

You build things people love to use. Whether it’s a workflow, a tool, or a feature, you focus on making it intuitive, efficient, and actually helpful.

You’re both a team player and an independent thinker. You thrive in collaborative environments but don’t need hand-holding to take charge of your work.

You vibe with our values—RECIPE. Resilience, empathy, camaraderie, integrity, play, and excellence aren’t just buzzwords—they’re how we work and what we celebrate.

£12 per hour, subject to applicable PAYE/NI deductions

Length of internship: 3 months (with possibility of extending) from January to March 2025

This may be a part-time or full-time role, and may be office-based, remote, or hybrid.

If you are based in London, we encourage you to work from our office in Clerkenwell at least 2 days a week. We believe there's a lot to learn and gain from being together!

You'll report to our Engineering Lead.

Candidates must have the right to work in the UK.

Stage 1: Submit your CV and fill in the application form below.

Stage 2 (60-min video call): You'll meet with one of our engineers, and we'll chat about your interests, experience and past projects. This is a technically-focused call.

Stage 3 (Final 45-min interview): We'll invite you to meet with another member of our team (our CEO) for a culture fit interview.

Success! If all goes well, we'll make you an offer conditional upon reference checks.