Data Scientist

The role We're building the analytical backbone behind how we prioritize leads and drive them toward conversion. Early analysis has already surfaced strong signals in the data, and we now need someone to take this from early-stage analysis to a properly validated model that the business can actually run on. In this role, you'll be involved in end-to-end stages: jointly framing the problem, statistical modelling, validating it rigorously, deploying and shaping the output into something that genuinely drives value on ground. A glimpse of what you’ll build A sharper way to prioritize • Build a lead prioritization engine that not only identifies high- & low-potential leads, but eventually also bring conversion predictability across lead cohorts • Replace heuristic/rule-based scoring with statistically sound model — chi-square, information value, logistic regression and other relevant approaches A real model of what drives conversion • Identify which factors in the sales process genuinely move conversion • Build a structured, data-backed model to identify what patterns, actions and combinations drive maximum conversions Keeping it alive • Dynamic model which continuously updates with additional information, time decay and with chain of actions • Regular revalidation as new conversion data accumulates Where this is headed: as this scales, the work grows toward A/B-tested rules, uplift modeling, and ML-based models. The analytical/statistical models will setup the early foundation for the same as we gain more depth and insights from early deployments What skills and experience you’ll bring to the table Must-have skills & experience • Logistic regression — coefficients, odds ratios, significance testing, Wald tests • Chi-square, Information Value, Cramér's V, WoE • Detecting multicollinearity, noise vs. real effect • Python (pandas, numpy, scipy.stats, scikit-learn) • SQL • Strong Google Sheets/Excel skills as a parallel tool • 2-3 years of experience in building regression-based or scoring models • Time-series decay function design • Dashboarding - Looker, Tableau, Data Studio Good to have • Exposure to scoring/scorecard work — e.g., credit risk, risk scoring, lead scoring • A/B testing or experiment design basics • NLP for signal extraction from unstructured data • Sales CRM familiarity (Salesforce, HubSpot, or similar) • Conceptual awareness of causal inference or uplift modeling Who you are • Can explain logistic regression to a Sales Head without jargon • Push back when data doesn't support the wanted conclusion • Care whether the model changes behaviour, not just whether it's correct • Good at documenting assumptions and model specifics to ensure the model can be audited and improved over time.