Headline

Building Comify.io | Senior Data Scientist | Machine Learning | IIT Kanpur | Python | Demand Forecasting | Retail CPG | Pharma Marketing | Credit Risk | Time Series | Statistics | AWS | ex-Quantiphi, Axtria, Spring, Aria

About

With ~7 years of experience in Machine Learning Engineering and Applied Data Science, I have designed, deployed, and scaled ML solutions that power credit decisions, fraud detection, demand forecasting, and recommendation systems across FinTech, Retail, and Pharmaceutical domains. * Recently at Aria Intelligent Solutions (Client: Unilever), built demand forecasting pipelines using Nixtla TimeGPT on Databricks using PySpark, modeling multivariate time series with exogenous drivers for global markets to support data-driven supply chain planning. * At Spring Financial (Canada), enhanced credit risk models and loan underwriting workflows using AWS Lambda, GitHub Actions, and MLflow, improving model accuracy and automation efficiency. * At Axtria, built content recommendation and route optimization systems, leveraging PyConcorde, OR-Tools, and SHAP to drive data-driven marketing and sales optimization. * Previously at Quantiphi, I architected a multi-stage hierarchical forecasting engine for a leading beverage client in North America, managing 1M+ time series across 300K stores, reducing forecast error by 20%, and improving supply chain planning. I migrated legacy models to Azure Databricks, built distributed PySpark pipelines for large-scale automation, and led a team of ML Engineers, driving consistent delivery quality and client satisfaction. * Experienced in Python, PySpark, SQL, AWS, and Azure Databricks. Across all roles, I have demonstrated a strong ability to bridge business and technical goals, collaborate cross-functionally with product, engineering, and policy teams, and deliver production-grade ML systems that enhance credit decisions, customer experience, and operational efficiency—fully aligned with the values and mission of any ML / AI organisation.

Location

Gurugram

Work experience

Senior Data Scientist at Comify, Sep 2026 - Present · 2 mos, Gurugram, Haryana, India, Building the Recommendation Engine at Comify.

Senior Data Scientist at Unilever, Dec 2025 - Aug 2026 · 9 mos, New Delhi, Delhi, India, Handling projects for Unilever as a Data Science contractor via Aria Intelligent Solutions. * Implemented demand forecasting pipelines using TimeGPT (Nixtla) on Databricks (PySpark clusters) with exogenous drivers for US and Poland markets; processed ~50K IDs (~10M data points) to enable data-driven supply chain planning. Improved forecast accuracy by 5% and reduced prediction bias by 1% through model tuning and feature enhancements.

Senior Data Scientist at Aria Intelligent Solutions, Nov 2025 - Aug 2026 · 10 mos, New Delhi, Delhi, India, Handling projects for Unilever as a Data Science contractor via Aria Intelligent Solutions. Majorly involved in enhancing their forecasting tools and modules for better forecasts over various regions across US and Poland. 1. Implemented demand forecasting pipelines using TimeGPT (Nixtla) on Databricks (PySpark clusters) with exogenous drivers for US and Poland markets; processed ~50K IDs (~10M data points) to enable data-driven supply chain planning. 2. Improved forecast accuracy by 5% and reduced prediction bias by 1% through model tuning and feature enhancements. 3. Designed and executed controlled experiments on Databricks across TimeGPT variants, fine-tuning strategies, and feature configurations, improving generalization across demand patterns. 4. Integrated external regressors (promotions, seasonality, datetime features) to improve forecast stability. 5. Conducted segment-level analysis across product/category hierarchies to capture diverse demand variability.

Machine Learning Engineer 2 at Spring Financial, Aug 2024 - Sep 2025 · 1 yr 2 mos, Gurugram, Haryana, India, 1. Credit Risk Modeling & MLOps: 🔹 Enhanced the Credit risk ML model for personal loans by integrating business-specific features to improve risk profiling and loan approvals; developed using VS Code and deployed via AWS Lambda and GitHub Actions. 🔹 Pioneered retraining capabilities for the delinquency ML model, achieving ~20% higher F1 score, 300% increase in true positives and ~ 30% fewer false positives. 🔹 Integrated advanced feature engineering using OpenAI APIs to extract semantic context from unstructured text, improving model signal and downstream credit risk classification. 🔹 Delivered a POC for MLOps using MLflow, demonstrating experiment tracking and model versioning for improved workflow transparency and reproducibility. 🔹 Learned and applied Airflow concepts, understanding orchestration, scheduling, and DAG management for future pipeline scalability. 🔹 Improved model performance by ~10% through experimentation with XGBoost, Random Forest, and class imbalance techniques (SMOTE, undersampling, SMOTEENN). 🔹 Implemented automated testing with Pytest across ML workflows, embedding CI-friendly quality checks to ensure robustness and consistency in production. 2. Data Ingestion & Pre-processing Pipelines: 🔹 Extended financial data ingestion pipeline by integrating Flinks, Plaid and TransUnion APIs, leveraging AWS Lambda for real-time automated preprocessing, reducing manual overhead and seamlessly supporting new data providers. 🔹 Developed a large-scale ETL & pre-processing pipeline in Python on AWS SageMaker to process ~10M records from Snowflake, automating SQL, cleaning, deduplication, and sampling. 🔹 Implemented multithreading for parallelized transformation, cutting processing time and manual effort by over 60% and delivering consistent ML-ready data. 🔹 Collaborated with business stakeholders and data engineers to ensure alignment of technical design with compliance, operational, and analytical requirements.

Senior Data Scientist at Axtria - Ingenious Insights, Nov 2023 - Jul 2024 · 9 mos, Gurugram, Haryana, India, 1. ML Modeling & Advanced Analytics: 🔹 Collaborated with cross-functional teams, including external stakeholders, to architect, develop, and deploy ML and statistical models for pharma and marketing analytics use cases, delivering actionable, data-driven recommendations. 🔹 Analysing patterns and trends in the data to identify actionable insights through rigorous EDAs, visualizations, Cohort analysis and Causal Inference in a fast-paced, dynamic environment. 2. Key Projects: 🔹 Content Recommendation System: Designed a content recommendation system for Healthcare Professionals (HCPs) based on engagement data across promotion channels, driving improved NBRx conversion and engagements for targeted HCPs. 🔹 Route Optimization POC: Developed an optimization engine using PyConcorde, OR-tools, Graphhopper and routingpy; visualized the route using Folium for planning the most prioritized and economical route for sales reps. 🔹 Studied the interaction of user-item and item-user matrix and calculated similarity using Cosine and Gower similarity for mapping similar users together. 🔹 Explored model interpretability using SHAP for feature importance and transparency. 🔹 Introduced model evaluation frameworks and documented best practices for reproducibility.

Education

Bachelor of Technology - BTech, Engineering, INDIAN INSTITUTE OF TECHNOLOGY KANPUR