Artificial Intelligence Engineer in Manchester

Ready to level up your career? The AI team at Airtime is looking for its next great addition. Airtime isn't just another rewards platform, we're a data-driven powerhouse reshaping how millions of people save money and how the world's biggest brands connect with their customers through seamless everyday shopping. We turn transactional data into actionable insights using industry leading technology. By blending analytics with pure creativity, we help retail giants truly understand their customers, boosting loyalty and driving revenue along the way.

Why Airtime?

• We're growing fast: We constantly launch slick new features to keep our massive member base and brand partners smiling.
• We're obsessed with innovation: It's what keeps us ahead of the pack.
• We crush goals together: We hold ourselves to high standards, but we win as a team.

If you're ambitious, collaborative, and ready to make a real impact, you'll fit right in.

What You'll Be Responsible For:

• Build relevant search: Design, ship and iterate on semantic search end-to-end using embeddings, vector similarity, hybrid keyword/semantic retrieval and reranking, so every query returns genuinely relevant offers.
• Blend ML relevance with commercial reality: Build ranking systems that combine model scores with retailer eligibility, offer availability, popularity, commercial rules and personalisation.
• Own ML systems end-to-end: Data preparation, training and embedding pipelines, deployment, monitoring, retraining and model/version management. Your systems will be production-quality, typed and tested.
• Deliver low-latency services: Design fast inference and retrieval services that hold up at scale, and integrate them cleanly via APIs with our customer-facing applications.
• Prove what works: Build evaluation into everything: model performance, failure analysis and online A/B testing, so we optimise for real customer and commercial impact.
• Develop our GenAI capabilities: Use and extend our shared LLM tooling for classification, enrichment, structured outputs and evaluation, and help deliver roadmap capabilities such as retrieval-augmented generation (RAG), agents and conversational experiences.
• Keep the platform healthy: Operate and improve our MLOps and cloud infrastructure: ZenML and Vertex AI pipelines, Docker, CI/CD and Terraform on GCP, with an eye on reliability, security and cost.
• Make every pound count: Choose models and architectures by weighing quality, latency and cost, apply FinOps practices to keep our cloud and LLM spend visible and under control, and make sure features earn their keep.
• Provide technical leadership: Partner with Data, Product, Marketing and Tech to turn ambiguous problems into measurable outcomes, make pragmatic build-vs-buy calls, and help shape our engineering standards.
• Above all, we want somebody pragmatic: an engineer who can own the full journey from data and experimentation through to a reliable product feature, and who knows when a simple exact-match search, taxonomy-based approach or managed service beats a complex vector database.

Experience:

• Strong Python and software-engineering skills, with experience writing production-quality, typed and tested code. We use pytest, mypy and Ruff.
• Experience owning and productionising ML systems on a major cloud platform.
• Strong SQL skills and experience working with large analytical datasets.
• Strong applied data-science experience, including statistical modelling, feature engineering, model fitting, validation and performance assessment across supervised and unsupervised learning problems.
• Practical experience in search, information retrieval, ranking or recommendation systems, combined with a rigorous approach to search evaluation: labelled query–retailer datasets, recall@k, precision@k, MRR/NDCG, relevance judgements.
• Experience in online experimentation (A/B testing, contextual bandits) to assess model or product performance.
• The ability to design and operate reliable, low-latency inference or retrieval services for customer-facing applications at web scale. Comfortable with latency budgets, caching, monitoring and alerting, and graceful degradation under real traffic.
• Experience making quality/latency/cost trade-offs in production ML or LLM systems: selecting models, right-sizing infrastructure & reducing spend without degrading quality.
• A track record of owning technical delivery from first experiment to production feature.
• Experience raising engineering quality through activities such as code review, knowledge sharing, the effective use of coding assistants, and setting standards.
• You do not need to have used every technology in our stack. We are more interested in strong engineering judgement, relevant experience and the ability to learn quickly.

Bonus Points If You Have:

• Improving embedding / ranking models through hard-negative mining, contrastive learning, fine-tuning or distillation.
• Learning-to-rank, candidate generation, collaborative filtering, implicit-feedback modelling or personalisation.
• Working with behavioural search data such as impressions, clicks and conversions, including position-bias-aware evaluation.
• LLM fine-tuning and practical LLM applications such as classification, labelling, and structured outputs.
• Retrieval-augmented generation (RAG), including ingestion, chunking, grounding, citations and retrieval evaluation.
• Query understanding: making short, misspelt or vague searches return relevant results via spelling correction, synonyms, taxonomy mapping, zero-result handling.
• Building AI agents or chatbots with appropriate guardrails, evaluation and observability.
• GCP, particularly Vertex AI, BigQuery, Cloud Storage or Pub/Sub.
• ZenML, Docker, CI/CD, dbt, and infrastructure-as-code tooling such as Terraform.
• Advanced experimentation or causal-inference techniques, such as power analysis, uplift modelling or treatment and holdout design.
• Experience working with high-volume behavioural or transactional data to build customer-facing models, ranking systems or decisioning products in domains such as retail, e-commerce, offers or fintech.

What You Will Get:

We believe great work happens when people feel supported, trusted, and genuinely valued. So, we've built a benefits package that's designed to give you flexibility, support your wellbeing, and help you grow both inside and outside of work.

Reward & Security:

• Share options: Be a true part of our success.

Time Off That Matters:

• 23 days holiday, plus more with every year you're here (up to 28!)
• Your birthday off: Go ahead, have a 'you day'
• Buy extra holiday when you need it (up to 5 extra days)
• A dedicated charity day to give back to causes you love.

Flexibility & Balance:

• Flexible start times: Roll in anytime between 6:30–10:30am.
• Hybrid working to perfectly balance home and office life.
• Private medical insurance.
• Health cash plan for everyday healthcare.
• Virtual GP access for you and your family.
• 24/7 support for mental health, counselling, and wellbeing.

Growth & Development:

• Learning & development budget + dedicated time to actually use it.
• Professional accreditation funding.
• Real opportunities to learn from experienced colleagues and level up your career.

Support Through Life:

• Enhanced maternity, paternity & adoption leave.
• A culture that understands life happens outside of work.

Community & Culture:

• Company charity contributions.
• Regular team moments that actually feel meaningful (no "forced fun" here).
• A team that values collaboration, curiosity, and just getting things done.

The Bottom Line:

We're building a place where you can do your best work without burning out, standing still, or feeling like just another cog in the machine. Because when you're supported, everything else follows.