Data Scientist , AI & Data Lead

About Natural Intelligence

Natural Intelligence is building AI products that turn complex, continuously evolving data into useful knowledge, predictions and decisions for a global market.

We are looking for an Applied Data Scientist who will lead the data work end to end: shaping the data architecture, establishing quality standards, uncovering meaningful relationships, and turning validated discoveries into production AI capabilities.

This is a hands-on technical leadership role. You will set the direction for the data layer, make and implement technical decisions, and be accountable for its quality and performance. You will work directly with our founder, infrastructure team and frontend/design team.

We are building a globally facing enterprise product, and the data foundations, analytical rigour and production reliability must meet that ambition.

What you will leadData strategy, architecture and quality

• Own the technical direction and delivery of the data layer, from raw ingestion through analysis and production use.
• Design data models and pipelines for extremely large, live, multimodal datasets, connecting structured records, documents, images, audio, video and sensor streams through consistent identities, timestamps and metadata.
• Establish standards for data quality, provenance, versioning and reproducibility, including robust handling of missing, inconsistent, delayed and duplicate data.
• Partner with infrastructure engineering on scalable storage, streaming, processing and access, and identify what additional data should be collected to improve the product.

Applied research and discovery

• Find meaningful connections across datasets and translate them into testable hypotheses and product opportunities.
• Lead statistical investigation and experimentation using methods such as forecasting, time-series analysis, classification, clustering and anomaly detection.
• Distinguish correlation from causal evidence, quantify uncertainty and validate promising research methods against practical baselines.
• Recommend which opportunities to pursue, what evidence is needed and when an approach should be stopped.

Production AI and data capabilities

• Turn validated findings into predictive models, data services, recommendations and intelligent workflows, including data foundations for retrieval, agent tools and multimodal AI.
• Own evaluation methods, acceptance criteria and monitoring for data and model quality.
• Partner with infrastructure and frontend colleagues to deliver enterprise capabilities that meet accuracy, latency, throughput, cost, privacy and access requirements.
• Monitor data and model drift, investigate failures and improve systems using real outcomes and user feedback.

AI-native, agentic ways of working

We build AI products, and we use AI extensively to build them.

You are expected to use coding agents and agentic workflows throughout research, data exploration, pipeline development, experimentation, testing and documentation. You should know how to break complex work into tasks, provide agents with appropriate context and tools, and validate their outputs.

As the data lead, you will also establish repeatable AI-assisted workflows that improve the team’s speed and quality.

You remain accountable for analytical correctness, engineering quality and the integrity of every result. Strong technical judgment and rigorous verification are essential.

Required background and experience

• At least 5 years of relevant professional experience in applied data science, statistical modelling, data engineering or closely related work, including leadership of a substantial data workstream or platform.
• A strong academic foundation in Statistics, Applied Mathematics, Operations Research or a related quantitative discipline, with depth in probability, statistical inference, linear algebra, optimization and experimental design.
• Strong Python and SQL, with experience writing maintainable, tested production code and processing large-scale, streaming or continuously updated datasets.
• Experience connecting structured and unstructured data, and delivering data or ML capabilities from messy source data through evaluation into production use.
• Sound judgment in model evaluation, including data leakage, overfitting, uncertainty and drift, supported by evidence of making architectural decisions and owning delivery.
• Established use of AI coding agents and agentic workflows in profess…