Applied Scientist (AI Evaluation)

Why join us? At Trismik we are a team of tech geeks from the University of Cambridge, Salesforce, and Amazon looking to push the boundaries of AI through science-led evaluations. If you are ambitious to make a difference to the future of AI, have a PhD in NLP, and like to turn ideas into reality, we would love to hear from you.

Role: We are developing adversarial tests for Large Language Models (LLMs). We are looking for a passionate, talented, and innovative applied scientist with a strong background in algorithms to help build an industry-leading evaluation engine for LLMs and help us bring this to market. As one of our first hires, this is a high-impact and high-ownership role. You will have a strong say in how our science and product roadmap evolves.

Our mission is to provide the fastest and most accurate testing environment to add value to AI engineers wishing to deploy AI applications using LLMs. We do this to push forward the state-of-the-art in Artificial Intelligence and to achieve the best possible chance of a human-aligned AGI.

Key job responsibilities: As an Applied Scientist in a startup, you will have a key role in our team. You will work with the Chief Scientist to design, develop and deploy evaluation technologies that will create high-value insights for AI engineers. These will involve providing assessments for several technologies that involve Large Language Models, including retrieval augmented systems (RAGs), recommender systems, and agentive systems. You will:

• Invent, experiment with, and launch new features, products and systems based on machine learning and MLLM.
• Perform hands-on construction and analysis of large-scale multi-modal datasets to be deployed as part of our product offering.
• Build algorithms, perform offline and A/B test experiments, optimise and deploy your models into production, working closely with software engineers.
• Establish automated processes for large-scale data analysis and generation, machine-learning model development, model validation and serving.
• Communicate results and insights to both technical and non-technical audiences, including through presentations and written reports, and publish your work for internal and external audiences, e.g. through white papers and conference papers.
• Collaborate with engineers, product partners locally and abroad, and join conversations with customers led by our marketing team.

Essential skills:

• PhD in Computer Science or related field (Natural Language Processing focus only).
• Experience in state-of-the-art deep learning models architecture design and deep learning training and optimisation and model pruning.
• Experience with LLM benchmarking.
• Strong programming skills in Python, Java, C++, or a related language.
• Strong skills in algorithm development to solve optimisation problems.
• Familiarity with core tools for a typical ML-focused work environment (e.g. Git, IDE, common libraries, HuggingFace).
• Ability to work as part of a team.
• Be comfortable with a fast-paced, high-risk, and high-reward environment.

Desirable skills:

• Experience creating data sets.
• Experience coordinating development of products with multiple stakeholders.
• Publications in top-tier peer-reviewed journals and conferences (e.g. NeurIPS, EMNLP, ACL).
• Experience deploying solutions to AWS or other cloud platforms.
• Excellent communication skills, solid work ethic, a strong desire to write production-quality code.

Our offer: Competitive salary, bonus, and share options in our startup; Company issued laptop, workstation, and tech allowance; Flexible/remote work environment. We are based in Cambridge/London but open to outstanding remote work candidates based in the UK or EU. We offer travel allowance to enable meet ups.

Contact: For informal enquiries please contact nigel@trismik.com.