Arham Mallick

– Optimized month-end payroll processing pipelines for legacy platforms by bulk-loading datasets and centralizing computation logic, driving a time complexity reduction from O(N×Q) to O(N+Q) for highly scalable operations – Architected a real-time notification subsystem leveraging Server-Sent Events (SSE) and environment-driven mailers to ensure reliable, concurrent delivery of mission-critical project and HR alerts – Secured platform infrastructure by isolating candidate portals onto strictly host-gated

- Created LLM-based scripts to tag and normalize 2,000+ manually entered site comments, reducing manual review effort significantly. - Engineered solar power monitoring algorithms using PySolar radiance data, identifying up to 47 kWh/day in wasted energy. - Developed an underutilization detection system for solar power that highlighted opportunities to reduce energy costs by up to 35%. - Designed and deployed an internal Flask web app with secure auth and real-time reporting, serving 100+ internal users.

- Built and deployed ML models to classify sizing factors (battery, rectifier, genset) with up to 98% accuracy. - Automated alarm insights by building Python pipelines integrated with a live web portal, enabling real-time diagnostics.

- Worked with multilingual text datasets on Google Cloud Platform, supporting translation alignment and quality review. - Extracted and structured content from web sources to build and refine annotation-ready corpora. - Annotated 2,000+ images using Roboflow for object detection tasks involving environmental monitoring scenarios. - Contributed to the training pipeline of a YOLOv8-based model by ensuring high-quality, consistent labeling standards.