Puranjay Yadav — AI Researcher

I am a Data Scientist with a Master’s in Data Science from Stevens Institute of Technology and professional experience at Sephora, Jaguar Land Rover, NVIDIA, and Robert Bosch. My expertise spans Machine Learning, AI, Statistical Modeling, and Big Data Architecture, with a proven track record of building scalable solutions that deliver measurable business impact. At Sephora, I developed and deployed predictive & recommender models (Wide & Deep, Two-Tower, PinSage) and built ETL/ML pipelines in Python, R, PySpark, Databricks, Snowflake, BigQuery, and Vertex AI, improving CTR by +12% and forecasting accuracy by +5%. At Jaguar Land Rover, I applied time-series forecasting (ARIMA, SARIMAX, XGBoost) and causal modeling to optimize inventory and sales, boosting sales velocity by 20% and reducing unsold stock by 25%. During my AI research with NVIDIA, I built deep learning models (YOLOv7, ResNet, RetinaNet) for supply chain analytics and real-time decision-making, improving accuracy by 20%. At Robert Bosch, I engineered ETL pipelines, anomaly detection systems, and automated visual inspection using computer vision, improving throughput by 35% and cutting inefficiencies. I specialize in connecting technology with business impact by leveraging AI, MLOps, Cloud Platforms (Azure ML, GCP, Databricks) and Business Intelligence (Power BI, Tableau). My passion lies in applying explainable AI (SHAP, causal inference), experimentation, and statistical analysis to uncover insights, drive personalization, and optimize decision-making across industries.

Stackforce AI infers this person is a Data Scientist with expertise in AI and machine learning across retail, automotive, and robotics industries.

Location: Jersey City, New Jersey, United States

Experience: 2 yrs 8 mos

Skills

Machine Learning
Generative Ai
Data Science
Business Analytics
Data Analysis
Full-stack Development

Career Highlights

Proven track record in building scalable AI solutions.
Expertise in machine learning and data-driven decision making.
Significant improvements in business metrics across multiple industries.

Work Experience

SEPHORA

GenAI Scientist (2 mos)

JLR North America

Product Data Scientist (11 mos)

Stevens Institute of Technology

Graduate Research Assistant (9 mos)

Bosch

Machine Learning Engineer (6 mos)

Full-stack Developer (6 mos)

WeddingsOnly

Junior Web Developer (5 mos)

Education

Master's degree at Stevens Institute of Technology

Bachelor's degree at Manipal University Jaipur

High School Diploma at Manav Sthali School - India

Puranjay Yadav

AI Researcher

Jersey City, New Jersey, United States2 yrs 8 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Proven track record in building scalable AI solutions.
Expertise in machine learning and data-driven decision making.
Significant improvements in business metrics across multiple industries.

Stackforce AI infers this person is a Data Scientist with expertise in AI and machine learning across retail, automotive, and robotics industries.

Contact

work.puranjay@gmail.com LinkedIn

Skills

Core Skills

Machine LearningGenerative AiData ScienceBusiness AnalyticsData AnalysisFull-stack Development

Other Skills

Advanced JavascriptArduinoAsync JavascriptAzure DatabricksBashBusiness KnowledgeC++Cascading Style Sheets (CSS)ChromaDBCommunicationCompetitive AnalysisCompetitive IntelligenceContinuous Integration and Continuous Delivery (CI/CD)Cross-functional PartnershipsCustomer Relationship Management (CRM)

About

Experience

2 yrs 8 mos

Total Experience

9 mos

Average Tenure

Current Experience

Sephora

GenAI Scientist

Jun 2025 – Aug 2025 · 2 mos · San Francisco Bay Area · Remote

Researched and implemented recommender systems (dual-encoder, GPT-based ranking, deep neural nets) to improve personalized discovery and search; delivered a +12% CTR lift.
Led data source identification & quality checks across Snowflake/BigQuery and NoSQL catalogs (null/outlier profiling, schema drift, SLA conformance) to ensure training data fitness.
Built large-scale feature engineering pipelines in Spark/PySpark; designed embeddings and behavioral features for ranking and RAG-powered content recommendations.
Defined model assessment & validation (offline ROC-AUC, PR-AUC, RMSE; online A/B), analyzed feature impact, and iterated using error slicing.
Deployed models via MLflow-tracked CI/CD (Docker, FastAPI); added monitoring for latency, throughput, and drift to sustain production performance.
Built interactive dashboards in Tableau/Python to track recommendation KPIs; collaborated with PMs to translate insights into go-to-market and backlog prioritization; followed WCAG 2.2 AA-aware practices.

Deep LearningMachine LearningGenerative AIRecommender SystemsPythonAzure Databricks

Jlr north america

Product Data Scientist

Jun 2024 – May 2025 · 11 mos · Mawah · On-site

Executed big data analysis and causal inference to identify CSAT/NPS drivers, informing lifecycle touchpoints and contributing to +15% NPS.
Built predictive models (Random Forest, Gradient Boosting) to forecast Days-to-Sell, improving inventory turnover +15% and sales velocity +20%.
Designed ETL/feature workflows in PySpark/SQL; created analytics-ready datasets and a lightweight feature store
for cross-functional reuse.
Established model validation protocols (cross-validation, ROC-AUC, PR-AUC, MAE/RMSE) and guided feature selection with permutation importance/SHAP.
Translated business questions into tech problem formulations; prioritized requirements and SLAs; supported data strategy/governance

Competitive AnalysisReporting & AnalysisBusiness AnalyticsDigital TransformationDatabase DesignTroubleshooting+16

Stevens institute of technology

Graduate Research Assistant

Jan 2024 – Oct 2024 · 9 mos · Hoboken, New Jersey, United States · On-site

Developed object detection/tracking (YOLOv7, ResNet, SSD, RetinaNet), improving accuracy +20% and robot navigation success +7%.
Built ML pipelines in TensorFlow/Keras; practiced rigorous experiment design, cross-validation, and dataset versioning
for reproducibility.
Performed quantitative evaluations (precision/recall, ROC/PR curves, latency trade-offs) and synthesized industry insights for stakeholders.
Authored technical/strategic reports translating model results into partnership and deployment opportunities.

Data ResearchCustomer Relationship Management (CRM)Statistical Data AnalysisData Science

Bosch

2 roles

Machine Learning Engineer

Promoted

Jan 2023 – Jul 2023 · 6 mos · Jaipur, Rajasthan, India · On-site

● Performed Data Analytics for Process Audits. Refined and arranged data in an organized manner to perform Analytics upon it, recognized the pattern and correlation between different features and used ML techniques to predict and prevent anomalies in components
● Designed and created an auto visual inspection system (AVIS), an AI system for pump calibration line to remove manual checking of OK seals of the pump. Used Single shot Multibox algorithm as Deep learning model along with other Image processing techniques to achieve quality deliverables on time
● Developed an Automated Mobile Robot by creating custom ROS modules along with customized Unscented Kalman Filter and embedded PID controller into the system. Managed a team of 5 to provide quality deliverables on time.

Robot Operating System (ROS)Data AnalysisTableauPython (Programming Language)Embedded CMachine Learning

Full-stack Developer

Jan 2022 – Jul 2022 · 6 mos · Jaipur, Rajasthan, India · On-site

● Automated the entire workflow from software to hardware and contributed to the value stream.
● Provided smart and efficient solutions to two major problems: removing manual labour and reducing the time required to study the analytics of different pumps and complaints in the Quality Management Methods department.
● Developed a system that automatically asks auditing questions for the parts received from the customer, Reduced paper storage and cost.
● Developed applications in Node.js, Django, JADE, MongoDB, and SQL.

React.jsNode.jsMongoose Web ServerDjangoJavaScriptOpenShift+3