Karachi Division, Sindh, Pakistan
- Led design and architecture of multi-agent AI platforms using GPT-4/5 and LangChain, reducing manual workflows. - Delivered enterprise-grade RAG pipelines with Pinecone and PostgreSQL, improving document query accuracy by 65%. - Deployed HIPAA-compliant chat and voice AI agents used by 50+ healthcare practitioners across the US. - Supervised a remote engineering team of 4 to ensure high-performance, production-grade deployments.
- Led a cross-functional team of 4 engineers to deliver an AI-driven invoice processing system that reduced vendor approval time by 35%. - Developed AI-powered automation frameworks for logistics, procurement, and finance verticals - reducing client manual data processing by 50–60%. - Built and deployed FastAPI-based microservices handling high-volume process automation. - Integrated LLMs into CRMs, Twilio, and Zapier, automating customer interactions. - Improved workflow automation through optimized LangChain prompt pipelines. - Developed an NLP-based document classification and entity extraction model using GPT-3 and spaCy to automate contract analysis, improving accuracy by 40% compared to rule-based methods.
- Developed NLP modules for document classification and summarization with 92% accuracy using TensorFlow. - Created REST APIs to manage document insights and reduce data retrieval time - Contributed to backend optimization efforts that improved overall application response time = - Collaborated with 3 cross-functional teams to deliver end-to-end software solutions for enterprise clients.
- Designed and implemented REST APIs and automation tools improving client reporting speed. - Supported AI module integrations into enterprise dashboards, reducing repetitive workflows. - Enhanced data reliability and security through database optimization and error logging systems. - Contributed as a full-stack developer on internal ERP and analytics applications using React (frontend) and Flask (backend). - Developed a workflow automation platform for a logistics client that digitized shipment tracking, resulting in 30% faster operational turnaround.
- Assisted the R&D department in developing OCR preprocessing pipelines for Arabic and English text. - Contributed to document classification tasks achieving ~88% accuracy in internal model benchmarks. - Helped build a text extraction and labeling utility in Python to accelerate dataset preparation, reducing annotation time. - Collaborated with senior engineers on NLP experiments using TensorFlow, Keras, and NLTK for entity recognition and keyword tagging. - Conducted research and performance comparisons between rule-based vs. CNN-based OCR models, producing a report that guided future system enhancements.