Senior Data Scientist

Role & responsibilities

• Mine, process, and analyse hit/event level web, product, sales, and digital marketing data. 
• Leverage LLMs (Large Language Models) and traditional machine learning to mine, process, and analyze web, product, sales, and digital marketing event-level data.
• Develop and fine-tune LLM-driven solutions for tasks such as text summarization, customer support automation, personalization, and user journey understanding.
• Build and deploy predictive models and ML algorithms across structured and unstructured customer profile, journey, and usage datasets.
• Deploy LLM and ML models into production environments for activation across websites, product applications, and sales/marketing channels.
• Design and implement model activation strategies, including A/B testing plans, benchmarking studies, and measurement of final business impact.
• Conduct comprehensive evaluation of LLMs, including performance benchmarking (accuracy, latency, token usage, cost), prompt effectiveness testing, fine-tuning impact analysis, and safety/bias assessments.
• Design, build, and deploy LLM-based agentic systems using frameworks such as LangChain, AutoGen, CrewAI, or custom orchestration for complex workflows (e.g., multi-agent collaboration, function-calling pipelines, dynamic task execution).
• Integrate LLM agents with APIs, internal knowledge bases, retrieval systems (RAG architectures), and external tools to enable autonomous or semi-autonomous decision-making.
• Partner with data engineering teams to enhance and maintain the Customer360 data model, including creating new feature engineering requirements, improving taxonomy, and identifying and resolving data quality issues.
• Collaborate with cross-functional teams (Enterprise Data Warehouse, Salesforce MOPS, IT, Product, Marketing) to continuously improve data integration and quality for advanced modeling use cases.
• Build a deep understanding of business models, objectives, challenges, and opportunities by working closely with leadership and key stakeholders.
• Document model methodologies, evaluation frameworks, agent workflows, deployment architectures, and post-activation performance results in a structured and reproducible format.
• Stay current with advancements in LLMs, agentic AI, retrieval-augmented generation (RAG), and ML technologies to recommend and implement innovative solutions.