Generative AI and retrieval
Builds RAG question-answering and agentic analytics workflows with embeddings, vector indexing, semantic search, prompt versioning, structured logging, and multi-provider model integrations for production use.
Senior Data Scientist & GenAI Specialist
Senior data scientist with 7+ years delivering AI and machine-learning systems across retail, e-commerce, fintech, and agriculture. Specializes in generative AI, RAG, NLP, voice systems, forecasting, and MLOps. Experience spans LangGraph orchestration, document retrieval, demand models, speech recognition, computer vision, and cloud deployment.
Pakistan
Builds RAG question-answering and agentic analytics workflows with embeddings, vector indexing, semantic search, prompt versioning, structured logging, and multi-provider model integrations for production use.
Develops sales and demand models, including low-history SKU methods, similarity mapping, proxy-series alignment, regressor-based forecasting, feature engineering, and anomaly detection across retail analytical workflows.
Applies speech recognition, entity extraction, and visual models to order-taking, document detection, identity verification, and crop disease classification with specialized deep-learning and NLP frameworks.
Turns exploratory notebooks into automated workflows and production deployments using Dataiku, Python, PySpark, AWS, GCP Dataproc, Docker, and operational integration practices for scalable AI systems.
Leads GenAI, document retrieval, agentic analytics, and demand forecasting work, integrating model APIs and orchestration with production workflows, prompt versioning, logging, Dataiku automation, and cloud-based MLOps practices.
Developed sales forecasting, sales-impact models, anomaly detection, feature engineering, and GCP MLOps workflows for retail analytics, focusing on demand prediction, data quality, processing efficiency, and production system operation.
Built voice-based order automation for Domino's branches, developing speech recognition and entity extraction pipelines with model-serving optimizations, containerization, and Linux operations to support scalable conversational order-taking workflows in production.
Developed AI components for KYC and fraud detection, including document detection and in-house face verification, using Python and computer vision frameworks to improve processing efficiency and reduce external service dependency.
Developed deep-learning and image-processing components for crop disease detection, supporting an AI advisory platform with Python model development and data processing for agricultural disease identification and prevention workflows.
Highlights from professional projects, product delivery, and specialist work.
Implemented retrieval-augmented document question answering with embeddings, vector indexing, and semantic search. Integrated model APIs with structured logging, prompt versioning, and LangGraph or custom orchestration, supporting reproducible inference and internal analytical use across retail workflows.
Built a voice-based chatbot for restaurant order-taking across 70+ US branches. Combined Kaldi and Wav2Vec speech recognition, SpaCy entity extraction, Docker deployment, and ZeroMQ optimizations for scalable, low-latency conversational interactions and structured ordering workflows.
Developed retail forecasting systems at Walmart and Visionet, including Prophet regressors, feature engineering, anomaly detection, and low-history SKU techniques. Automated model workflows on GCP Dataproc and Dataiku to move exploratory work into production analytics.
FAST National University of Computer and Emerging Sciences
Aug 2021 – Jun 2023
University of Management and Technology
Sep 2015 – Jun 2019
Builds document-grounded and tool-using AI systems with attention to reproducible inference, orchestration, logging, production workflows, and integration with business analytics and data platforms.
Combines forecasting, anomaly detection, speech, NLP, and computer vision to address retail demand, conversational ordering, identity verification, and agricultural disease-detection requirements.
Share your requirements, current challenges, and delivery goals. Botmer will review the brief and help arrange the next step.