SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Chattermill is a Customer Experience Intelligence platform that helps major brands (Uber, Amazon, Wise, HelloFresh, etc.) analyze customer feedback using advanced AI and machine learning. The company builds custom ML models specialized for customer feedback tasks across extraction, retrieval, reranking, summarization, and sentiment analysis.
You will join the Data Science team to build and ship the next generation of their ML stack. Your responsibilities will include:
- Train, evaluate, and iterate on ML models for customer feedback tasks, contributing to custom fine-tuning pipelines and running rigorous experiments with clear documentation.
- Build and maintain LLM-powered features including retrieval pipelines, reranking systems, and insight generation, with support from senior team members.
- Contribute to evaluation frameworks by building test sets, defining metrics, and assessing model quality across classification, extraction, and generative tasks.
- Work on semantic search and retrieval, developing expertise in embedding-based approaches and methods beyond standard techniques.
- Write clean, well-tested code and collaborate with Engineering on model integration, data pipelines, and monitoring.
- Translate business and product requirements into practical ML experiments and solutions alongside the wider Data Science team.
- Stay current with relevant research and bring useful ideas into team discussions and experiments.
The company emphasizes pragmatism: the right solution may combine off-the-shelf LLMs, bespoke fine-tuned models, or non-LLM techniques as appropriate.
Chattermill offers flexible working (remote or hybrid), 25 days holiday plus bank holidays, annual learning budget, equity options, health and wellbeing support, and a dog-friendly London office with rooftop terrace.
REQUIREMENTS:
- Solid working knowledge of transformer architectures and their application in NLP tasks
- Proficiency in PyTorch, including training loops and standard model fine-tuning workflows
- Experience working with real-world text data at meaningful scale (classification, extraction, embeddings, or search)
- Some exposure to instruction fine-tuning or model serving, with interest in going deeper
- Grounding in classical ML and statistics; instinct to use simpler methods when appropriate
- Familiarity with GenAI and agentic patterns
- Clear communication skills and ability to explain technical work across functions
- Genuine curiosity about AI and habit of experimenting
- Good ownership instincts and follow-through on problems
BONUS:
- MSc in Computer Science, Machine Learning, AI, Data Science, Computational Linguistics, or related STEM field
- Exposure to parameter-efficient techniques such as LoRA
