Jagjot Bisram

Organized AI workshop series (PyTorch, Hugging Face, FastAPI) with 50+ attendees; guided members in building working ML prototypes such as chatbots, image classifiers, and REST APIs. Implemented project management frameworks (sprints, GitHub workflows, reproducibility standards), boosting project completion rates from 40% to 80%. Led team project delivering a loan approval prediction web app with XGBoost + FastAPI, mentoring students on class imbalance handling, model evaluation, and REST API deployment.

Built predictive fault detection models on sensor time-series data, reducing equipment downtime by 10%. Automated data preprocessing workflows with Pandas, cutting runtime by 30% and accelerating model iteration. Partnered with data engineers to integrate models into real-time production pipelines, enabling proactive maintenance alerts. Conducted exploratory data analysis on sensor streams to identify leading indicators of equipment failure, informing feature selection. Documented model performance and pipeline workflows for handoff to operations team.