Innovative Machine Learning Engineer with a Master’s in Network Science, specializing in NLP and AI-driven solutions. Creator of the largest Persian hate speech dataset, with expertise in cutting-edge AI systems for real-world applications. Passionate about advancing the frontiers of machine learning and impactful research.
Integrated a Retrieval-Augmented Generation (RAG) based question-answering system into the company’s customer chatbot, improving query understanding, response accuracy, and overall customer support capabilities.
Developed and implemented Liveness Detection algorithms for e-KYC systems, incorporating facial recognition, motion analysis, and biometric data to verify user authenticity and enhance security.
Enhanced Face Verification accuracy by optimizing the FaceNet algorithm implementation, resulting in a significant improvement in performance from 91.32% to 98.34%.
Utilized Docker to enhance e-KYC projects, ensuring consistent environments, streamlined dependency management, and easy deployment. Improved scalability, resource management, and CI/CD integration, fostering collaboration and productivity.
Developed an LDA-based text summarization system for call transcripts, expediting the identification of customer pain points and boosting call center efficiency, while integrating real-time analysis and CRM connectivity to deliver actionable business insights.
Improved sentiment analysis algorithms for call transcripts, achieving a substantial increase in customer satisfaction detection accuracy, from 83.2% to 91.1%.
Developed a predictive model to estimate personnel exit times, enabling proactive resource planning and reducing employee turnover.
Improved demand forecasting accuracy using advanced machine learning algorithms, optimizing inventory, and operations while reducing stock-outs and costs across industries.
Designed and implemented comprehensive HR solutions to manage the needs of over 800 personnel, providing essential reports for HR department and management teams.
Developed a comprehensive data warehouse solution with efficient ETL processes, integrating data from existing ERP systems to ensure data integrity, accessibility, and centralized data management.
Implemented powerful BI reports using Power BI, enabling stakeholders to gain insights through interactive visualizations and data-driven decision-making and process optimization across the organization.
Created BI dashboards for CEO-level HR analytics, including workforce turnover, payrolls, and work hours, enabling data-driven decision-making and strategic planning.
Developed a Project Management system integrated with Microsoft Project Server, focusing on Business Intelligence solutions to enhance project planning and tracking.
Designed and implemented a Security Management System integrated with Biostar attendance devices, aimed at effectively monitoring and controlling user traffic within the organization.
Developed a Quality Assurance Management system utilizing suitable technologies to meet specific demands, ensuring high-quality deliverables and efficient processes.
Implemented tailored timesheet and tasksheet solutions to accommodate the diverse needs of employees across departments, generating crucial reports for HR and management purposes.
Led recruitment efforts, designing and implementing rigorous software interview questions and code challenge assessments.
Successfully hired and mentored a team of 5 developers from top Iranian universities, facilitating their professional growth and knowledge enhancement.
Contributed to the development of a Persian Medical Question Answering System in a research lab, applying advanced NLP techniques for accurate information retrieval and response generation to improve access to critical medical information for Persian speakers.
Designed a novel Crowdsourcing-based Open Innovation Business Model leveraging collaborative problem-solving and idea generation to drive organizational growth.
Instructed PyTorch workshops for students encountering challenges with dissertation projects, promoting effective deep learning implementation and research progression.
Instructed introductory Python courses for new students in faculty forums, fostering foundational programming skills and engaging the academic community.
Facilitated simulation and modeling course through homework correction and unwind sessions, supporting students’ understanding of key concepts and applications.
Created and taught a neural network lessons, including plans, unwind sessions, and homework correction, fostering an engaging learning environment for students.
H. Davardoust, H. Zare, H. Rafiee Zade, The Dark Side of Instagram: A Large Dataset for Identifying Persian Harmful Comments” submitted to SoCal NLP Symposium 2024 (Southern California Natural Language Processing Symposium), (Accepted, to appear).
H. Rafiee Zade, H. Zare, M. Ghassemi Parsa, H. Davardoust, M. Shariat Bagheri, ”DCOR: Anomaly Detection in Attributed Networks via Dual Contrastive Learning Reconstruction,” Thirteenth International Conference on Complex Networks and Their Applications 2024, (Accepted, to appear)
H. Davardoust, H. Zare, ”Complex network analysis of keywords co-occurrence in Graph Neural Network, Learning on Graphs Conference 2024” (under review)
Ranked within the top of 0.5% of the Iranian University Entrance Exam for Master’s degree amoung 43,532 students