Data scientist and researcher specializing in machine learning, time-series forecasting, and large-scale system optimization.

My work focuses on bridging theoretical modeling with real-world applications in content delivery networks (CDNs), streaming systems, and data-driven infrastructure. I am particularly interested in efficient algorithms, predictive modeling, and energy-aware system design.

I have contributed to multiple research projects combining statistical modeling and deep learning:

• Developed a hybrid LSTM–attention architecture for short-term forecasting of live TV viewership, integrating temporal dynamics, contextual features, and regime-aware modeling. The model achieves highly accurate predictions on large-scale IPTV data

• Designed a Markov-chain-based analytical model for Perfect LFU caching, addressing a long-standing gap in modeling frequency-based cache algorithms and enabling practical CDN optimization and sizing decisions

• Proposed Top-LRU, a novel cache eviction strategy combining neural forecasting with traditional caching methods, improving cache hit ratios by up to 7.2% in live streaming environments

My work sits at the intersection of:

– Machine Learning (LSTM, forecasting, neural architectures)

– Data Engineering & large-scale data analysis

– Algorithm design and stochastic modeling

– Distributed systems and streaming infrastructure