Open to opportunities

Karthikeya Pervela

@karthikeyapervela

Data Scientist specializing in end-to-end ML systems, recommendation engines. I build scalable, production-ready solutions.

India

Message

What I'm looking for

I am looking for a role that fosters innovation and collaboration, where I can apply my data science skills to solve real-world problems and contribute to impactful projects.

Data Scientist with 2+ years of experience building production-grade ML systems, from feature engineering to A/B testing and deployment. At EloElo, I built a personalized short-video recommendation system from scratch, delivering +1.85% D1 retention lift via online A/B experiments. I engineered features using PySpark on EMR Serverless, built an online feature store (ScyllaDB) for sub-50ms inference, and containerized the scoring API for AWS ECS. I replaced p-value reporting with bootstrap-based lift probability for faster decisions. Previously at Saint-Gobain, I developed cost-sensitive XGBoost models on IoT sensor data, improving defect recall to 84% and reducing unacceptable batches by 14%. I also built a constraint-programming optimization system that improved material utilization to 95% and cut waste by 50%. I hold an M.Tech from IIT Kanpur, with expertise in ML, statistics, and optimization. Passionate about scalable, data-driven solutions for complex business challenges

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Eloelo

Sep 2025 - Present (10 months)

Built a personalized short-video recommendation system from scratch—designed custom loss functions, engineered behavioral features with PySpark on EMR, deployed a sub-50ms scoring API with ScyllaDB, and validated +1.85% retention lift through online A/B experiments. Replaced p-value reporting with bootstrap-based lift probability for faster decision-making.

Python AWS Lambda SparkSQL Lambda EMR Docker fastAPI CursorAI Claude SQL Recommendation Systems Recommendation Engines Recommender Systems AB Tests Google Sheets Data Pipelines Feature Stores (Tecton Scylla

Data Scientist

Freelance Projects

Aug 2023 - Apr 2024 (8 months)

Developed a demand forecasting system and a financial fraud detection engine, utilizing advanced machine learning techniques and cloud deployment for real-time applications.

Gradient Boosting SQLite Cost Sensitive Learning AWS EC2 Object Oriented Design

Data Scientist

Saint Gobain

Aug 2023 - Apr 2024 (8 months)

Led ML-driven materials research optimization initiatives for R&D teams, optimizing workflows and enhancing model accuracy. Collaborated with cross-functional teams to develop actionable ML solutions.

Python Machine Learning Cross Functional Collaboration Data Analysis AWS EC2 SQL

Education

Degrees, certifications, and relevant coursework

Indian Institute of Technology, Kanpur

Master of Technology, Machine Learning

2021 - 2023

Focused on the application of Machine Learning for materials property prediction, with relevant courses in Machine Learning, Mathematics, and Computational Methods.