Kruthika Kumar Muralidharan

Lead a 40-member analytics organization (including 4 people managers) supporting Marketing, Customer Success, and multiple product areas across Razorpay. Built self-serve measurement and decision systems, standardized experimentation and auto-diagnostics, and led production ML/GenAI initiatives including internal AI assistants, evaluation frameworks, and workflow automation.

I led the Sales, Delivery & Risk analytics teams (team size: 38), as well as the Data Science centre of excellence. My key focus areas here were: standardization of metrics (L1, L2) and key drivers for my business teams, automation & standardization of all dashboards on a central platform, and building predictive models (lead scoring & fraud detection). Key projects: Lead scoring: Led the creation of an ML model that improved identification of high-quality leads 10x. The insights from this exercise were also used to improve targeting of digital marketing spends.

Headed the Analytics, Planning and Product Strategy teams at Ecom Express, one of the largest e-commerce logistics players in India. Led 4 teams with 24 bright folks during my time here, helping solve difficult problems using data. Key responsibilities: Analytics: • Providing dashboards, reports & data-driven decision-support tools (e.g. pricing models) to internal functions • Undertaking longer term performance improvement projects with advanced data modelling 2. Planning: • Forecasting volumes & planning resources for the entire supply chain (2,800+ centres, and 1 million+ lanes), and working with operational teams to drive productivity improvement 3. Product Strategy: • Scoping, design & launch of new business products, including market scoping, process design, coordination with internal/external stakeholders, technology design & implementation Key analytics projects: i) Volume forecasting with a customised variant of exponential smoothing techniques ii) A locality identification/address correction solution to reduce incorrect routing that uses text and geospatial analytics iii) A simulation model to predict last mile supply chain failures in advance, which uses a machine learning (XGBoost) based model to predict productivity fluctuations, iv) a probabilistic weight-correction system that incorporates package weights from multiple sources (equipment, historical data etc), and assigns the correct weight to each package.

Worked in the transportation infrastructure space, consulting for private sector and government clients on a variety of projects, including infrastructure project planning & viability assessment (traffic forecasting, assessment of the operational and financial viability), business process reengineering, fund raising advisory, transaction advisory & business growth strategy. Work scope: I was the Engagement Manager for multiple projects, with responsibility for full project delivery. This role involves engagement planning, data collection and analysis, guiding junior members and output review, coordination with technical teams and external experts, output preparation, client meetings and presentations, billing, and collections. Business Development: Responsible for the preparation of proposals and client interactions at the business development stage.

Project in Tata Capital.

Area: Natural Language Processing (NLP), Deep learning Thesis research: Benchmarking of a variety of models (including transformer based architectures like multilingual BERT, XLM-R) on sentiment analysis of Hinglish code-mixed social media text.

Specialization: Deep Learning