Luqman Khan

Demand Management Model for New Opportunity

• Designed an Azure-native AI solution integrating Azure OpenAI, text embeddings, and Azure AI Search.
• Designed an Azure-native AI solution integrating Azure OpenAI, text embeddings, and Azure AI Search.
• Built a scalable Retrieval-Augmented Generation (RAG) system using GPT-4o and LangChain.
• Automated deployment via CI/CD pipelines in Azure DevOps, with model telemetry and monitoring.
• Collaborated with cross-functional teams to define requirements and align delivery with product goals.
  

LLM-Based Job Description Generator

• Engineered a GPT-4o-based system to generate customized job descriptions from demand templates.
• Reduced manual effort by 80% and supported an internal hiring tool across more than five business units.
• Used prompt engineering techniques to tailor responses based on domain and role context.
• Deployed the solution with a full DevOps pipeline, logging, and exception monitoring.
  

KEY DEEP LEARNING PROJECTS:

Melanoma Detection

• Built and augmented image datasets using Keras preprocessing.
• Developed a CNN model with dropout and batch normalization to improve robustness.
• Achieved high classification accuracy (92%) on early-stage melanoma detection.
  

Gesture Recognition for Smart TV.

• Designed and trained a real-time gesture recognition model for embedded edge devices.
• Focused on low-latency inference using a lightweight CNN architecture.
• Reduced false positives by 40% with improved training on augmented datasets.