Zulfiqar Ali

Architect, train, and optimize deep-learning models for real-time audio, video, image, and text analytics (speech-to-text, logo detection, sentiment classification, object tracking). Build scalable ingestion pipelines that handle multi-modal feeds from TV, radio, OTT, and social-media sources, then containerize the models (Docker/Kubernetes) and deploy them on GPU-enabled on-prem or hybrid cloud clusters. Establish CI/CD, versioning, rollback, and A/B testing workflows to ensure continuous improvement. Complement the deployment with interactive dashboards—built in Grafana, PowerBI, or React/D3—that surface live KPIs, anomalies, and compliance alerts for senior stakeholders.

As a Data Analyst at the Press Information Department under the Ministry of Information & Broadcasting, I built automated pipelines to collect and preprocess real-time data from diverse sources, including text, images, and videos, for machine learning applications. I developed real-time analytics dashboards using Power BI and Python to monitor media trends and model performance. My work involved multimodal data analysis, combining natural language processing and computer vision techniques to extract insights from multimedia content. Additionally, I applied statistical and machine learning methods to uncover patterns, engineer features, and support predictive modeling for public sentiment analysis and misinformation detection.

At Sejong University’s Center for Photonic Systems, I worked as a Graduate Research Student in the Department of Optical Engineering, where I applied artificial intelligence and advanced signal processing techniques to improve biomedical imaging for cardiovascular disease diagnosis. My research involved developing robust imaging pipelines and extracting features from micro-endoscopy data for predictive modeling. I also focused on engineering disease-specific imaging biomarkers to support precision diagnostics and enhance model interpretability. This experience deepened my interdisciplinary expertise across biomedical engineering, AI, and optical imaging systems.

Contributed to research and development at the National Center of Artificial Intelligence (SmartCity Lab), where I built high-dimensional probabilistic models for vision-based robot navigation and 3D scene reconstruction. Designed decision-making systems for autonomous agents operating in dynamic environments, and developed learning-based control strategies using pose estimation and motion planning. My work directly supported advancements in intelligent transportation and robotics.

During my tenure as a Graduate Research Assistant at COMSATS University Islamabad, I focused on statistical data science and machine learning applications to real-world behavioral datasets. My work primarily involved applying Hierarchical Bayesian modeling techniques to analyze and classify large-scale behavioral survey data. I designed and implemented semi-supervised learning pipelines to handle imbalanced and partially labeled datasets, ensuring model robustness and generalization. Additionally, I engineered end-to-end data preprocessing workflows incorporating Bayesian imputation and Expectation-Maximization (EM) algorithms to enhance model performance and handle missing data effectively. This experience significantly deepened my expertise in probabilistic modeling, advanced machine learning methods, and data-centric research

Project on Children's Behavior Classification Problem using Machine Learning

I have successfully completed the first two courses — Financial Data and Financial Markets — with a cumulative grade above 80%, making me eligible for certification Foundation of Financial Engineering