Nitin Patel

AI Researcher

Bengaluru, Karnataka, India4 yrs 2 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

Expert in designing production-grade ML systems.
Significantly improved search-to-cart conversions.
Passionate about data-driven decision-making.

Stackforce AI infers this person is a Data Scientist specializing in E-commerce and Machine Learning.

Contact

Skills

Core Skills

Machine LearningPredictive Modeling

Other Skills

SQLPythonData AnalysisStatistical Data AnalysisDBTProduct AnalyticsPrompt EngineeringFinancial ModelingDeep LearningComputer VisionC++CMicrosoft WordRubyCollege Football

About

With a Master of Science in Economics from the Indian Institute of Technology, Kharagpur, I currently work as a Senior Data Scientist at Zepto, where I specialize in designing and deploying production-grade machine learning systems. My core competencies include predictive modeling, SQL, and building scalable data pipelines, which I leverage to enhance search ranking, personalization, and user engagement. At Zepto, I contributed to the development of an intent-aware search ranking system, improving search-to-cart conversions significantly, and created a personalized replenishment prediction system that enhanced buy-again behavior. Passionate about leveraging data for actionable insights, I thrive on collaborating with cross-functional teams to empower data-driven decision-making and align technical solutions with organizational goals.

Experience

4 yrs 2 mos

Total Experience

1 yr 4 mos

Average Tenure

2 yrs 1 mo

Current Experience

Zepto

2 roles

Senior Data Scientist

Promoted

Apr 2025 – Present · 1 yr 1 mo · Bengaluru · On-site

I design and build production-grade machine learning systems for search and personalisation, driving product discovery, repeat purchases, and conversion at scale.
🔹 ML Systems & Modeling
Led end-to-end development of an intent-aware search ranking system, combining query intent classification and learning-to-rank models to dynamically optimize product ordering — improving search-to-ATC conversion from 37% to 50%.
Built a personalized replenishment prediction system using gradient boosting and probabilistic (Weibull) models to forecast user–SKU repurchase timing, enabling proactive recommendations and driving a 70 bps uplift in buy-again conversion.
Engineered large-scale feature pipelines capturing user behavior, lifecycle signals, seasonality, and query–product relevance to support robust model performance across sparse and high-frequency use cases.
Designed ranking models leveraging behavioral signals (CTR, ATCR), semantic relevance, and category alignment to optimize product discovery across diverse query intents.
🔹 Production Systems & Scalability
Productionized ML pipelines with automated feature generation, monitoring, and periodic retraining to ensure freshness of behavioral signals in a high-frequency consumer environment.
Built scalable data pipelines on distributed infrastructure (Databricks, Spark, Snowflake) to support real-time and batch inference workflows.
Developed experimentation frameworks and A/B testing pipelines to validate model performance and quantify business impact.
🔹 Product & Business Impact
Drove measurable improvements in search conversion, user retention, and repeat purchase behavior through ML-driven personalisation systems.
Partnered closely with product, engineering, and business teams to translate user behavior insights into deployable ML solutions aligned with growth, profitability and inventory objectives.