Farzaneh Lashgari

Designed and implemented large-scale data analytics pipelines for mental health datasets, enabling efficient processing and analysis of complex, multi-source conversational data. Developed an ensemble-based suicide risk detection system leveraging three LLMs (transformer-based architectures) combined with advanced feature engineering, improving detection accuracy from ~80% to 93%. Engineered a multi-agent AI framework for automated detection and response generation, reducing detection and response latency t

• Developed advanced analytics solutions for HR and financial systems, improving decision-making speed by ~35% • Built automated data pipelines, reducing manual workload by 40%+ and significantly improving data accuracy • Designed interactive dashboards (Tableau, SQL), enabling real-time business monitoring • Identified inefficiencies in cost structures, contributing to data-driven cost optimization strategies Key Project: HR Analytics & Compensation Optimization • Designed models to optimize compensati

• Led development of data-driven solutions for 60K+ employees, one of the largest workforce datasets in Iran • Built large-scale data pipelines for workforce analytics and operational optimization • Developed executive dashboards supporting strategic decision-making Key Project: Intelligent Payroll & Overtime System (Big Data) • Designed dynamic salary/overtime model based on performance + attendance + productivity metrics • Reduced ~20,000 hours of non-productive overtime • Increased workforce efficie

• Started as Data Analyst and progressed to IT Manager role • Managed IT infrastructure, network operations, and system reliability • Designed reporting systems and analytical tools for internal departments • Built strong foundation in data, systems, and infrastructure integration • Led cross-functional collaborations between technical and operational teams

designing intelligent and adaptive detection models. I developed a hybrid malware detection framework that combined both supervised and unsupervised learning approaches in a unified architecture, enabling more robust identification of both known and previously unseen threats. The system leveraged

development, working extensively with Pascal, C, and C++. As part of my academic training, I developed several small-scale software applications and games, strengthening my understanding of algorithms, data structures, and problem-solving. For my final project, I designed and implemented a reverse