Senior Data Scientist

Humara; Brighton; Hybrid; GBP 90,000 - 130,000; Full time

Benefits offered by this job

Medicash healthcare scheme (dental, -)

Private Medical Insurance via AXA (1st

25 days holiday + bank holidays + your

Employee Assistance Programme

Gogeta nursery salary sacrifice

Enhanced parental leave and pay

Job summary

Humara is hiring a Senior Data Scientist to lead hands-on development of LLM-powered multi-agent sales technologies. You will design and deploy production-grade AI that drives conversion, working with data scientists, AI engineers and MLOps specialists in cross-functional teams.

You’ll build retrieval pipelines, manage model training and evaluation, and ensure dependable systems with testing, observability and governance.

Qualifications

• Extensive ML/DS experience with production-grade AI systems.
• Hands-on NLP, LLMs and agent-based AI development.

Responsibilities

• Design, build and ship LLM- and agent-based systems from prototype to live product.
• Build retrieval pipelines (RAG, ranking, summarisation, grounded Q&A).
• Train, fine-tune and optimize language models for production use.
• Define success metrics via offline evals, experiments and monitoring.

Skills

LLM systems

PyTorch

LangChain

Agentic AI

MLOps

Tools

Docker

AWS

LangGraph

Langfuse

Vector search

Job description

At Humara we're developing the next generation of virtual sales assistants. With genAI at our core, we help users discover products that truly fit their needs, provide Product Experts to answer their questions 24/7, and guide them through complex sales processes in any channel they choose.

We're looking for a Senior Data Scientist to join our established team and be a hands‑on contributor to new and existing features of our multi‑agent conversational sales product, with a focus on agentic and LLM‑driven systems that perform reliably in production and drive conversion. You'll work alongside a multi‑skilled group of data scientists, AI engineers and MLOps specialists in cross‑functional teams. We deliver against quarterly OKRs, own our work end to end, and put collaboration at the heart of how we operate.

Requirements

Strong ML experience, hands‑on NLP and production experience with agentic AI are essential. Our stack includes AWS, Python, PyTorch, Docker, LangChain/LangGraph and Langfuse. Experience with reinforcement learning or an understanding of sales psychology would be desirable plus there's plenty of ambitious work now and upcoming.

The person we're looking for

• Enjoys tackling hard, open‑ended problems and seeing the solution through to real users
• Has a get‑things‑done mindset and takes work from idea to delivery
• Keeps learning and stays current in a fast‑moving field
• Practical and collaborative: argues a position with conviction, and changes it when the evidence says so
• Explains technical ideas clearly to engineers and commercial teams alike
• Works independently and manages their own time and priorities well

You’ll be responsible for

• Designing, building and shipping LLM and agent‑based systems, from first prototype to live product
• Building retrieval pipelines (RAG, ranking, recommendation, summarisation, grounded Q&A)
• Training, fine‑tuning and optimising models, mainly for language tasks
• Defining how we measure success, through offline evals, experiments and production monitoring tied to user and commercial outcomes
• Keeping systems dependable, with automated testing, observability, troubleshooting and incident response
• Setting the bar for code quality, security, data governance and documentation
• Working closely with engineering, data and product teams, and mentoring others

Skills and experience

• Extensive experience in data science and ML, including hands‑on delivery of LLM‑powered and agentic systems running in production at scale
• Works AI‑natively day to day, using AI tools and agents as a core part of how they build, test and ship
• Strong Python, plus PyTorch, scikit‑learn or similar frameworks
• Hands‑on LLM work: prompting, fine‑tuning, grounding and guardrails
• Agentic systems at scale, including orchestration (LangChain/LangGraph), tracing and evals (Langfuse), memory design, and latency and cost trade‑offs
• Embeddings, vector search and hybrid search
• Solid grounding in statistics and ML theory across supervised and unsupervised learning, NLP (entities, intent), recommendation, prediction and uplift modelling
• Full model lifecycle and MLOps: containerisation (Docker/ECR) and deployment on AWS
• Working with large text datasets and data pipelines
• Nice to have: sales psychology applied in data products

Please note that this is a hybrid role and will require regular travel to our Brighton office.

Studies have shown that women and people who are disabled, LGBTQ+, neurodiverse or from ethnic minority backgrounds are less likely to apply for jobs unless they meet every single qualification and criteria. We're committed to building a diverse, inclusive, and authentic workplace where everyone can be their best, so if you're excited about this role but your past experience doesn't align perfectly with every requirement on the Job Description, please apply anyway - you may just be the right candidate for this or other roles in our wider team.

• Medicash healthcare scheme (reclaim costs for dental, physiotherapy, osteopathy and optical care)
• Private Medical Insurance via AXA (after 1 year service) Launching in April 2026
• 25 days holiday + bank holidays + your birthday off (rising to 28 after 3 consecutive years with the business & 30 after 5 years). Including Christmas closures.
• Employee Assistance Programme (confidential counselling)
• Gogeta nursery salary sacrifice scheme (save up to 40% per year)
• Enhanced parental leave and pay including 26 weeks' full maternity pay and 8 weeks' paternity leave