{ "name": "Nikunj Kaushik", "headline": "Analytics & Data Science Leader", "tagline": "Analytics, Data Science and AI", "location": "Bengaluru, India", "years_experience": "9+", "current": { "company": "Razorpay", "domain": "Fintech / Payments", "focus": "Growth & Marketing Analytics" }, "summary": "Analytics and data science leader with 9+ years turning analysis into business outcomes - revenue, retention and growth, not dashboards for their own sake. Mechanical and Automation Engineering graduate who started out in analytics. Now also builds AI tools that give business teams direct access to their data without an analyst in the loop for every question.", "email": "info@nikunjkaushik.com", "links": { "website": "https://nikunjkaushik.com", "linkedin": "https://www.linkedin.com/in/nikunjkaushik/", "github": "https://github.com/kaushiknikunj1", "x": "https://x.com/NikunjKsk" }, "experience": [ { "company": "Razorpay", "domain": "Fintech / Payments", "note": "Growth & marketing analytics", "current": true }, { "company": "MPL", "domain": "Mobile Gaming", "note": "Revenue, product and marketing data science" }, { "company": "Ola", "domain": "Mobility / Logistics", "note": "Supply planning analytics" }, { "company": "Mu Sigma", "domain": "Analytics Consulting / Fortune 500", "note": "Decision frameworks for CPG, pharma and technology enterprises" } ], "expertise": [ "Uplift modeling and causal inference", "Churn prediction and retention forecasting", "Multi-touch attribution", "Bayesian marketing mix modeling", "Revenue simulation and P&L forecasting", "Product, growth, marketing and FP&A analytics", "LLM and RAG systems for analytics", "Text-to-SQL" ], "stack": [ "SQL", "Python", "R", "Power BI", "Tableau", "R Shiny", "Google Cloud", "BigQuery", "Gemini API", "Advanced Excel" ], "available_for": [ "Freelance and fractional analytics / data science consulting", "AI and LLM analytics build-outs", "Senior analytics and data science roles" ], "generated": "2026-10-31T18:30:08Z", "pages": [ { "name": "Home", "url": "https://nikunjkaushik.com/index.html", "description": "Bio, experience, skills" }, { "name": "Projects", "url": "https://nikunjkaushik.com/projects.html", "description": "Eight built-with-AI projects (all code public) and six applied ML case studies" }, { "name": "Blog", "url": "https://nikunjkaushik.com/blog.html", "description": "Long-form writing on analytics, data science and AI" }, { "name": "AI Tools", "url": "https://nikunjkaushik.com/ai-tools.html", "description": "AI analytics product being built in the open" }, { "name": "Second Brain", "url": "https://nikunjkaushik.com/second_brain.html", "description": "Personal knowledge vault with a Gemini query layer and a graph view" }, { "name": "Contact", "url": "https://nikunjkaushik.com/contact.html", "description": "Get in touch" } ], "ai_projects": [ { "title": "Meridian - an autonomous agent", "url": "https://meridian.nikunjkaushik.com", "summary": "Reads the open web, keeps its own notes, chooses when to wake, reports one-way to Telegram; no tool can post, email or edit", "code": "https://github.com/kaushiknikunj1/meridian" }, { "title": "My Second Brain", "url": "https://nikunjkaushik.com/second_brain.html", "summary": "Interactive knowledge graph of everything written and built (321 nodes), plus an Obsidian vault with a Gemini Q&A CLI" }, { "title": "Analytics AI", "url": "https://nikunjkaushik.com/ai-tools.html", "summary": "Upload a CSV or connect a database, ask in plain English, get a chart and a straight answer (Gemini on Cloud Run; in progress)" }, { "title": "Ask my blog", "url": "https://github.com/kaushiknikunj1/ask-my-blog", "summary": "RAG with citations over 27 articles and case studies, index built from this site's llms.txt; retriever first, generator optional", "code": "https://github.com/kaushiknikunj1/ask-my-blog" }, { "title": "Churn Explainer", "url": "https://github.com/kaushiknikunj1/churn-explainer", "summary": "Per-customer SHAP drivers turned into two-line retention briefs; every LLM-written number verified against the model", "code": "https://github.com/kaushiknikunj1/churn-explainer" }, { "title": "Anomaly Narrator", "url": "https://github.com/kaushiknikunj1/anomaly-narrator", "summary": "Isolation Forest + robust z-scores find the spike; an LLM drafts the first incident note with every number checked", "code": "https://github.com/kaushiknikunj1/anomaly-narrator" }, { "title": "Text-to-SQL Analyst", "url": "https://github.com/kaushiknikunj1/text-to-sql-analyst", "summary": "Plain-English questions over a payments database with schema-aware prompts, read-only guardrails and one automatic repair", "code": "https://github.com/kaushiknikunj1/text-to-sql-analyst" }, { "title": "Agent-readable profile", "url": "https://nikunjkaushik.com/llms.txt", "summary": "llms.txt, profile.json, profile.md and RSS regenerated from the site by tools/gen_profile.py" } ], "case_studies": [ { "title": "User Churn Prediction Model", "url": "https://nikunjkaushik.com/projects/churn-prediction.html", "method": "Predictive ML", "code": "https://github.com/kaushiknikunj1/nikunjkaushik-demos/tree/main/churn-prediction" }, { "title": "Bayesian Marketing Mix Model", "url": "https://nikunjkaushik.com/projects/marketing-mix-model.html", "method": "Bayesian econometrics" }, { "title": "Uplift Modeling for Incremental Targeting", "url": "https://nikunjkaushik.com/projects/uplift-modeling.html", "method": "Causal inference", "code": "https://github.com/kaushiknikunj1/nikunjkaushik-demos/tree/main/uplift-modeling" }, { "title": "Multi-Touch Attribution Engine", "url": "https://nikunjkaushik.com/projects/multi-touch-attribution.html", "method": "Attribution modeling", "code": "https://github.com/kaushiknikunj1/nikunjkaushik-demos/tree/main/multi-touch-attribution" }, { "title": "Revenue Simulation & Forecasting", "url": "https://nikunjkaushik.com/projects/revenue-simulation.html", "method": "Monte Carlo simulation", "code": "https://github.com/kaushiknikunj1/nikunjkaushik-demos/tree/main/revenue-simulation" }, { "title": "Payment Metrics Anomaly Detection", "url": "https://nikunjkaushik.com/projects/anomaly-detection.html", "method": "Production ML" } ], "articles": [ { "title": "India 2003–2025 — UPA vs NDA, in 214 Charts", "url": "https://nikunjkaushik.com/blog/india-decade-in-charts.html", "category": "Geopolitics", "date": "Aug 21, 2026", "reading_time": "no reading required", "description": "The same two decades of India — UPA (2003–14) vs NDA (2014–25) — told entirely in 214 charts. No prose, no spin: every metric as a picture, every source linked in the companion article.", "work_in_progress": false }, { "title": "The CAC Payback Trap: Why Your Payback Period Is Lying to You", "url": "https://nikunjkaushik.com/blog/cac-payback-trap.html", "category": "Business & Finance", "date": "Jun 13, 2026", "reading_time": "8 min read", "description": "Why your CAC payback period is probably lying to you — blended vs paid CAC, gross-margin adjustment, and cohort-weighting, with a worked example.", "work_in_progress": false }, { "title": "LTV:CAC Is Lying to You", "url": "https://nikunjkaushik.com/blog/ltv-cac-truth.html", "category": "Business & Finance", "date": "Jun 11, 2026", "reading_time": "8 min read", "description": "The 3:1 LTV:CAC rule is quoted everywhere and computed correctly almost nowhere. Survivorship bias, discounting, and channel blending — with a corrected worked example.", "work_in_progress": false }, { "title": "Simulating the P&L Before You Make the Decision", "url": "https://nikunjkaushik.com/blog/simulating-the-pnl.html", "category": "Business & Finance", "date": "Jun 9, 2026", "reading_time": "9 min read", "description": "Most 'data-driven' decisions are justified after the fact. The real leverage is simulating the P&L before you commit — driver-based models, scenarios, and sensitivity analysis.", "work_in_progress": false }, { "title": "1929 → 2026 — The Next Great Reset", "url": "https://nikunjkaushik.com/blog/great-financial-reset-2026.html", "category": "Geopolitics", "date": "Jun 7, 2026", "reading_time": "15 min read", "description": "The structural parallels between 1929 and 2026 are uncomfortably close. A data-first look at the next great financial reset — the manufacturing reroute, and the case for an appreciating rupee.", "work_in_progress": true }, { "title": "India 2003–2025 — UPA vs NDA, by the Numbers", "url": "https://nikunjkaushik.com/blog/india-decade-comparison.html", "category": "Geopolitics", "date": "Apr 9, 2026", "reading_time": "40 min read", "description": "A data-driven comparison of India's two governments — UPA (2003–14) vs NDA (2014–25) — across 201 metrics spanning economy, infrastructure, health, education, space, security, manufacturing, and society. Verdicts based on absolute improvement and percentile-adjusted rankings.", "work_in_progress": false }, { "title": "Text-to-SQL with LLMs: What Actually Works", "url": "https://nikunjkaushik.com/blog/text-to-sql-llms.html", "category": "AI", "date": "Mar 4, 2026", "reading_time": "9 min read", "description": "I've built a few Text-to-SQL systems now. Here's what I got wrong the first time, what fixed it, and what most tutorials miss entirely.", "work_in_progress": false }, { "title": "LLMs for Analytics: Honest Take After a Year of Building", "url": "https://nikunjkaushik.com/blog/llms-for-analytics.html", "category": "AI", "date": "Mar 4, 2026", "reading_time": "7 min read", "description": "After spending a year integrating LLMs into analytics work at Razorpay and in side projects — here's what I actually think, without the hype.", "work_in_progress": false }, { "title": "RAG for Analytics: Letting Your Team Ask Questions in Plain English", "url": "https://nikunjkaushik.com/blog/rag-for-analytics.html", "category": "AI", "date": "Mar 3, 2026", "reading_time": "8 min read", "description": "Text-to-SQL solves the query problem. RAG solves the context problem — what 'activated' means in your product, why March is always lower, and all the institutional knowledge your LLM doesn't have.", "work_in_progress": false }, { "title": "How to Set Up Claude Code CLI: A Practical Guide", "url": "https://nikunjkaushik.com/blog/claude-code-setup.html", "category": "AI", "date": "Mar 1, 2026", "reading_time": "5 min read", "description": "A step-by-step guide to installing and configuring Claude Code CLI — Anthropic's AI coding assistant — for your development workflow.", "work_in_progress": false }, { "title": "Anomaly Detection Is the Easy Part. Explaining What It Means Is the Hard Part.", "url": "https://nikunjkaushik.com/blog/anomaly-detection-llms.html", "category": "AI", "date": "Feb 25, 2026", "reading_time": "8 min read", "description": "I built an anomaly detection system for payment metrics at Razorpay. The system worked. The 2am Slack alerts that said 'anomaly detected: 1.8σ' and left everyone frozen — that was the real problem.", "work_in_progress": false }, { "title": "Beyond Dashboards: Why AI is the New Interface for Product Analytics", "url": "https://nikunjkaushik.com/blog/ai-product-analytics.html", "category": "AI", "date": "Feb 21, 2026", "reading_time": "6 min read", "description": "How LLMs and AI are replacing static dashboards as the primary interface for product analytics teams. A practitioner's perspective.", "work_in_progress": false }, { "title": "Bayesian MMM in 2026: What's Changed, What Hasn't, and Where AI Fits", "url": "https://nikunjkaushik.com/blog/bayesian-mmm-2026.html", "category": "AI", "date": "Feb 14, 2026", "reading_time": "9 min read", "description": "I built a Bayesian MMM at MPL across 6 marketing channels. The model was good. Getting a CMO to act on adstock curves and posterior distributions — that was the actual problem.", "work_in_progress": false }, { "title": "Uplift Modeling: Measuring the True Impact of Your Campaigns", "url": "https://nikunjkaushik.com/blog/uplift-modeling.html", "category": "Data Science", "date": "Nov 18, 2025", "reading_time": "7 min read", "description": "Learn how uplift modeling moves beyond conversion rates to measure true incremental impact — and why it's the highest-leverage analytics investment for growth teams.", "work_in_progress": false }, { "title": "The Analyst's Guide to A/B Testing: From Hypothesis to Decision", "url": "https://nikunjkaushik.com/blog/ab-testing.html", "category": "Data Science", "date": "Oct 22, 2025", "reading_time": "8 min read", "description": "A practical framework for running rigorous A/B tests — from hypothesis design and sample sizing to interpreting results and making decisions under uncertainty.", "work_in_progress": false }, { "title": "Data Storytelling: Turning Analysis into Executive Action", "url": "https://nikunjkaushik.com/blog/data-storytelling.html", "category": "Product Analytics", "date": "Sep 5, 2025", "reading_time": "6 min read", "description": "How to communicate data insights in a way that drives decisions — structuring narratives, choosing the right visuals, and speaking the language of business.", "work_in_progress": false }, { "title": "Multi-Touch Attribution: Moving Beyond Last Click", "url": "https://nikunjkaushik.com/blog/multi-touch-attribution.html", "category": "Product Analytics", "date": "Jul 14, 2025", "reading_time": "7 min read", "description": "A deep dive into multi-touch attribution models — from linear and time-decay to data-driven approaches — and how to choose the right one for your marketing stack.", "work_in_progress": false }, { "title": "Churn Prediction with Machine Learning: A Practitioner's Guide", "url": "https://nikunjkaushik.com/blog/churn-prediction-ml.html", "category": "Data Science", "date": "May 3, 2025", "reading_time": "9 min read", "description": "An end-to-end guide to building churn prediction models with XGBoost and SHAP — from feature engineering to deployment and intervention design.", "work_in_progress": false }, { "title": "Choosing Your North Star Metric: A Framework for Product Analytics", "url": "https://nikunjkaushik.com/blog/product-north-star-metrics.html", "category": "Product Analytics", "date": "Mar 12, 2025", "reading_time": "6 min read", "description": "How to identify the single metric that best captures your product's core value delivery — and how to build your analytics strategy around it.", "work_in_progress": false }, { "title": "Cohort Analysis: What Aggregated Retention Hides", "url": "https://nikunjkaushik.com/blog/cohort_analysis.html", "category": "Product Analytics", "date": "Dec 27, 2024", "reading_time": "6 min read", "description": "How to use cohort analysis to understand retention, LTV, and the long-term impact of product changes on different user groups.", "work_in_progress": false }, { "title": "India / Bharat @ 2047 — A Civilisational Return", "url": "https://nikunjkaushik.com/blog/india-2047.html", "category": "Geopolitics", "date": "Coming soon", "reading_time": "Geopolitics", "description": "In 2047, India turns 100. What could the world's largest democracy achieve in the decade leading up to its centenary?", "work_in_progress": true } ], "feeds": { "rss": "https://nikunjkaushik.com/feed.xml", "llms_txt": "https://nikunjkaushik.com/llms.txt", "markdown": "https://nikunjkaushik.com/profile.md" } }