Web3 · AI · Product Analytics

Mohit Kumar

Business Analyst, Web3 & AI Systems

I work at the intersection of product, analytics and AI. I have shipped Web3 products, built marketing analytics systems, and trained machine learning models for forecasting and detection.

Email me +91-9304240766

accuracy in AI-vs-human content detection: 93%; accuracy in AI-vs-human content detection
better Google Ads tracking accuracy: 60–70%; better Google Ads tracking accuracy
higher customer engagement: 42%; higher customer engagement
precision in churn prediction: 98%; precision in churn prediction

Experience

Business Analyst (Web3 & AI Systems)
June 2024 – Present
Ravant Technologies (formerly Renske Tech)Onsite
- Delivered FOMO.gg (now ripr) within one month, owning it from research to deployment. It is a Solana-based platform where users create tokens in just two steps.
- Defined the product roadmap and turned stakeholder requirements into a scalable platform. Built marketing analytics in Tableau and Power BI covering retention, campaign performance and customer behaviour.
- Ran campaign measurement and ROI analysis across marketing operations. Implemented server-side GTM on Google Cloud and Stape across multiple products, improving Google Ads lead tracking accuracy by 60–70%.
- Owned end-to-end user lifecycle analytics: behavioural cohort frameworks, real-time dashboards and lifecycle health reports that surface dormancy signals, retention anomalies and campaign performance.
- Built reactivation scoring models (XGBoost, LightGBM) end to end, from feature engineering and cross-validation to score calibration and lift evaluation with gain charts, KS statistic and AUC, then turned scores into prioritised outreach lists for campaign tools.
- Built AI agents and chatbots across several use cases, and trained a RoBERTa model that detects AI vs human-written content at 93% accuracy, saving 600–800 USD a month in tool costs.
Product ManagementPythonAIWeb3Product AnalyticsMongoDBCohort AnalysisMachine Learning
AI/ML Specialist Intern
January 2024 – June 2024
Shaurya Infosoft (Partnered with Renske Tech)Onsite
- Developed and customised machine learning models for yPredict.ai, including a PyTorch LSTM with 84% accuracy for forecasting cryptocurrency prices from Yahoo Finance and other APIs.
- Improved customer engagement by 42% by explaining model insights and forecasts to non-technical audiences through clear documentation.
PythonMachine LearningStatistical ModellingTime Series Forecasting

Projects

Oct 2023 – Dec 2023

Predicting Customer Churn in a Subscription Service

Churn prediction model for subscription services using Decision Tree and Random Forest classifiers, reaching 98% precision through feature engineering and data exploration.

Machine LearningPredictive AnalyticsRandom Forest

Sep 2022 – Oct 2022

Advanced Text Analysis

NLP analysis beyond sentiment scores, giving insight for content creation, audience targeting and readability improvement.

PythonNLPRegular Expressions

Skills

Product

Product ManagementProduct OwnershipProduct StrategyRequirement GatheringStakeholder ManagementA/B Testing

Data & analytics

PythonSQLMongoDBTableauPower BICohort AnalysisTime Series Forecasting

AI & machine learning

Machine LearningPyTorchXGBoostLightGBMLogistic RegressionNLPLLMsAI Agents

Web3

SolanaSmart ContractsBlockchain FundamentalsWeb3 Product Architecture

Cloud & tooling

GCPAWSDockerCI/CDGitHubPostmanServer-side GTM

Ways of working

Cross-functional CollaborationData-driven Decision MakingAnalytical ThinkingCommunicationLeadership

Education

B.Tech, Computer Science and Engineering (AIML)
Noida Institute of Engineering and Technology, Greater Noida
Aug 2020 – Jun 2024CGPA 8.39
Class XII (BSEB)
Thakur Prasad Singh College, Patna
Apr 2017 – Apr 201981.2%
Class X (BSEB)
Bihar School Examination Board, Bihar
Apr 201774.6%