Developed and enhanced the AI detection tool at Orivfy, focusing on improving accuracy in text detection models. Implemented passage-level and sentence-level AI classification to refine model performance. Integrated explainable AI and provenance into product architecture, ensuring transparency and reliability.

Led the design and deployment of production-grade clinical AI systems, including MedScribe—a real-time medical transcription and report automation platform. Focused on building scalable, secure, and compliant AI pipelines for healthcare, integrating speech, NLP, and hospital systems like EHR and PACS. Managed system architecture, performance optimization, and close collaboration with clinicians for accurate and trustworthy AI solutions.

Contributed to development of an AI-powered medical imaging platform, working on DICOM viewers, computer vision pipelines, and multimodal AI systems. Combined medical imaging, language models, and backend services to support diagnostic workflows while ensuring scalability, security, and healthcare compliance. Documented system architecture and supported long-term maintainability for research and production use.

Built and deployed LLM-based legal intelligence systems tailored for Pakistani courts and legal professionals. Designed RAG pipelines, fine-tuned domain-specific language models, and developed semantic search systems for legal research and document drafting. Contributed to AI safety projects, including deepfake detection and audio cloning prevention, working across NLP, computer vision, and backend AI services.

Contract-based software engineering for federal education systems. Developed scalable solutions for educational technology infrastructure and data management systems.

Active contributor to open source AI/ML projects including model optimization, inference frameworks, and AI deployment tools.View on GitHub →