Analytics & Modelling
• Build, validate and deploy ML models for marketing use cases: audience segmentation, lead scoring and propensity, churn prediction, LTV estimation, demand/spend forecasting, and campaign attribution.
• Design and analyse A/B and multivariate experiments — define hypotheses, compute sample sizes, control for bias, and state results with confidence intervals rather than point estimates.
• Perform exploratory analysis on large marketing and web datasets; translate findings into explicit recommendations for client and internal stakeholders. Generative AI & RAG
• Build and improve retrieval-augmented generation pipelines: document ingestion, chunking strategy, embedding generation, vector storage and retrieval evaluation.
• Develop and tune prompts, evaluation harnesses and guardrails for LLM-backed assistants; measure quality with objective metrics (groundedness, retrieval precision/recall, latency, cost per request).
• Contribute to agentic and tool-calling workflows where LLMs invoke internal APIs and business tools. Data & Engineering
• Write production-grade Python; build reproducible feature pipelines and transformations over the lakehouse/warehouse.
• Own the full model lifecycle — versioning, offline/online evaluation, deployment, monitoring and retraining triggers; detect and respond to data and concept drift.
• Produce clear documentation: data dictionaries, model cards, assumptions, limitations and evaluation results. Collaboration & Governance
• Partner with engineering to expose models as APIs and with delivery teams to interpret results for clients.
• Apply privacy-by-design: PII minimisation, anonymisation/pseudonymisation, and handling of client data consistent with India’s DPDP Act and Singapore’s PDPA (several clients are Singapore-based).




• Never fabricate business numbers — cite data sources and state uncertainty explicitly. Required Qualifications
• 2–3 years of hands-on data science or applied ML experience in a product or client-delivery environment (internships and academic projects do not substitute for production experience).
• Bachelor’s or Master’s in Computer Science, Statistics, Mathematics, Data Science, or a related quantitative field.
• Python — strong proficiency with pandas, NumPy, scikit-learn; clean, testable, modular code (not notebook-only).
• SQL — comfortable with joins, window functions, CTEs and query optimisation over large tables.
• Solid grounding in statistics: hypothesis testing, regression, resampling, bias/variance, and the difference between correlation and causation.
• Practical experience with at least one of: gradient boosting (XGBoost/LightGBM), time-series forecasting, or NLP.
• Demonstrated experience with LLM APIs or frameworks — prompt design, embeddings, or a RAG implementation taken beyond a demo.
• Version control with Git and collaborative PR-based workflows.
• Clear written and spoken English; able to explain a model’s behaviour to a non-technical stakeholder. Preferred (Positive To Have)
• AWS experience — S3, Lambda, Bedrock, SageMaker, Athena, or Aurora PostgreSQL.
• Vector search in production — pgvector, OpenSearch, or similar, including retrieval evaluation.
• Spark / Databricks or equivalent distributed processing.
• MLOps tooling: MLflow, model registries, CI/CD for ML, containerisation.
• Marketing/adtech domain exposure — GA4, Meta/Google Ads data, HubSpot or other CRM data models, attribution modelling.
• Dashboarding and storytelling: Power BI, Streamlit, or similar.
• Open-source contributions, Kaggle results, or a public portfoli

Skills:
machine learning,sql,tableau,power bi,scikit-learn,python,pandas,statistical modeling

📌 Data Scientist (Coimbatore)
🏢 VenPep Group
📍 Coimbatore