Supply Chain Demand Forecasting
Developed LSTM-based time series model to predict product demand across 200+ SKUs. Reduced overstock by 34% and stockouts by 45%.
From concept to render — creating detailed 3D models for product design, mechanical engineering, and visualizations.
Designing reliable circuits and PCB layouts for embedded systems, power electronics, and custom IoT solutions.
Custom development board with ESP32-WROOM-32E, USB-C PD, battery management, and expansion headers.
FOC-controlled brushless DC motor driver with current sensing, over-voltage protection, and UART interface.
Low-power PIR motion sensor with tamper detection, RF connectivity, and AES-256 encryption.
Applying mathematical modeling and AI to solve real business problems — from demand forecasting to process optimization.
Developed LSTM-based time series model to predict product demand across 200+ SKUs. Reduced overstock by 34% and stockouts by 45%.
Built mixed-integer programming model to optimize job shop scheduling. Integrated with ERP for real-time replanning.
Created gradient boosting model to identify at-risk customers 60 days in advance. Enabled proactive retention campaigns.
Implemented reinforcement learning pricing system that adjusts prices based on demand elasticity and inventory levels.