Muhammed Shaheer Njezhukil

• Developed conversational assistants using LLMs, Vector embeddings, RAG - based frameworks to enable
dynamic, contextually relevant dialogue.
• Developed an Agentic AI workflow to automate aircraft supply operations by handling RFQ emails using
LangChain and LangGraph.

• Led as an ML Backend Engineer for a generative AI productivity suite, leveraging OpenAI chat models,
Azure Cognitive Search, and semantic similarity pipelines to accelerate DevOps workflows.
• Built a secure cloud repository for document management, enabling automatic classification, structured
storage, and fast search using Azure Cognitive Search across unstructured audio, video, and image files
with multi-label NLP; achieved 85% accuracy.
• Automated finance reimbursement and rig document processing workflows using Azure Form Recognizer,
streamlining document extraction and reducing manual effort.

• Built a transfer learning-based NSFW and low-quality image detection model, achieving 95% accuracy.
• Implemented the backend for a mobile app with an image similarity module using VGG16
• Deployed ML/DL models to AWS EC2 for scalable production use.
• Developed a real-time hand gesture recognition system using Mediapipe.
• Developed an Expert Recommendation System with text preprocessing