South Africa: Data Scientist AI/ML

Role Overview The company helps clients become data-driven by applying data technology to create real-time, actionable visibility into business performance and to build lasting analytics capability. This role will help expand delivery capacity for pragmatic, production-grade analytics, ML, and GenAI solutions, so client teams can move from “interesting insights” to deployed capabilities with measurable value. What youll do Translate client business questions into analytic approaches (metrics, segments, hypotheses), then deliver recommendations leaders can act on. Perform data exploration to identify drivers, anomalies, and performance constraints; communicate findings clearly to non-technical stakeholders. Design, run, and analyse experiments (A/B tests) and/or apply causal inference where experimentation is constrained. Build predictive models (e.g., risk/propensity/forecasting) with sensible baselines, validation, and error analysis, document tradeoffs. Build LLM/GenAI applications (prompt workflows, tool use, structured outputs) and implement RAG when retrieval improves accuracy. Create evaluation approaches for LLM systems (test sets, rubrics, automated checks, and human review loops) and iterate based on evidence. Implement guardrails: PII handling, secure prompting patterns, input/output validation, rate limiting, and fallback behaviors. Package models and GenAI components into clean services (APIs/batch jobs), with basic CI/CD awareness and production readiness. Improve/extend data pipelines and feature reliability (freshness, data quality checks, lineage-friendly design). Contribute to value measurement: define success metrics, track outcomes, and help clients understand what moved and why (the company emphasizes value measurement as part of delivery). Work within a structured engagement approach (Discover Review Implement Value Evolve), producing lightweight deliverables clients can run with. Role Expectations Strong fundamentals in SQL Python for analysis, transformation, and reproducible work (notebook production-friendly code). Experience delivering at least one end-to-end analytics or ML project (problem framing data model/analysis decision or deployment). Working knowledge of experimentation (A/B testing) and common pitfalls; comfort explaining statistical reasoning simply. Exposure to LLM application development (prompting, a small RAG prototype, or integrating an LLM into a workflow). ML engineering fundamentals: Git, code reviews, basic testing mindset, and ability to build/maintain a simple API or batch pipeline. Data quality mindset: you proactively validate inputs/outputs and can implement basic checks (freshness, nulls, ranges, duplicates). Responsible AI awareness: privacy/PII, security basics, bias considerations, and safe deployment practices. Communication skills: you can present to mixed audiences and write clear docs (assumptions, approach, results, recommendations). Comfort in a consulting-style environment: shifting context, collaborating with client teams, and managing expectations professionally. Nice To Haves Familiarity with cloud and modern data stacks used in consulting projects (warehouses and orchestration). Experience with BI/analytics enablement (semantic layers, dashboards, metric definitions, stakeholder training). Familiarity with vector search tooling and embedding workflows. Exposure to MLOps tooling (model tracking, monitoring, alerting) and basic incident/debugging habits. Any experience working with executives or operational leaders on decision-making and value measurement. Domain depth in a client-relevant industry: [finance/telecom/retail/public sector/health/etc.] Cloud experience (Azure/AWS/GCP) with any Data Science / AI / Data Engineering certifications. Databricks and/or Snowflake experience