Koushik Sundar

Algorithmic challenges involving NP-hard and NP-complete problems, such as a real-time subset sum problem with a sliding window, can be highly complex. One approach might be to reduce the subset size to keep polynomial computation time; however, with thousands of records entering and exiting the subset in real time, achieving optimal performance is challenging and can impact system efficiency. If you've faced this problem, did you experiment with any AI/ML or GenAI techniques to improve performance without compromising output quality? Alternatively, have you used AI/ML to guide the subset computation—perhaps by prioritizing or ordering calculations based on AI/ML-driven insights from historical data analysis? Or any other options being explored like quantum/edge computing as GPU is not going to make difference in this scenario.