Balaj Khalid

Data Science Intern

Alcon Inc.

Developed demand forecasting solution for US-based IOLs by building an end-to-end pipeline leveraging advanced time series models, driving data-backed inventory optimization and enhancing supply chain visibility; positioned to improve operational efficiency and generate significant cost savings.

AI Researcher

USC Center for Advanced Research Computing (CARC)

Conducted research on multi-agent reinforcement learning by implementing and comparing algorithms such as DQN, PPO, and coordination strategies like CTE, DTE, and CTDE; supplemented ongoing work in self-organizing systems to evaluate and improve the robustness and generalizability of reward functions across different learning models, contributing to a deeper understanding of scalable agent coordination.

Data Scientist II

Afiniti Ltd.

• Reduced model deployment time by 30% and enhanced operational efficiency by designing and automating a full-cycle MLOps pipeline, including raw data ingestion, transformation, model training, deployment, and monitoring.
• Enhanced productivity and efficiency by mentoring and leading a team of three in data analysis, predictive modeling, and production monitoring, fostering a high-performance culture.
• Achieved a 9% reduction in customer churn and generated €2M in additional revenue for Telefonica Spain by optimizing churn prediction models with advanced gradient boosting techniques (XGBoost, LightGBM)
• Increased data processing efficiency and contributed to €1M in incremental revenue by developing a real-time customer interaction tracking system using R, PostgreSQL, and automated ETL workflows with Talend. Developed predictive models for customer behavior analysis, created data pipelines, and collaborated with cross-functional teams

Data Scientist I

Afiniti Ltd.

• Generated €9.7M in additional revenue by successfully onboarding Telefonica Spain as a strategic client and designing real-time predictive modeling to optimize agent-caller pairing.
• Increased model interpretability and predictive accuracy, leading to $0.8M in additional annual revenue, by engineering domain-specific ML features tailored to business needs.
• Saved $0.5M in revenue by developing a real-time production monitoring and data integrity dashboard using Apache Superset, enabling proactive anomaly detection and preventing costly errors.

Master of Science in Computer Science

University of Southern California

Relevant course work: Artificial Intelligence, Machine Learning, NLP, Databases, Web Technologies, and Algorithms.

Bachelor of Science in Computer Science

Lahore University of Management Sciences

Relevant course work: Data Science, Artificial Intelligence, Machine Learning, Data Structures and Algorithms, Probability, Statistics, Linear Algebra, and Calculus.