Ahmar Kamran

PROFESSIONAL PROFILE Machine Learning/Computer Vision Engineer with a strong passion for AI, Large Language Models (LLMs), Experienced in developing keypoint detection models, image segmentation, and deep learning pipelines for real-world applications, training object detection models like fire and smoke detection, video analytics using opencv. Comprehensive understanding of 3D generation models (e.g., NeRFs), encoder architectures including Variational Autoencoders, Diffusion Models, GANs and data preprocessing techniques (PCA, LDA, UMAP) for dimensionality reduction and feature extraction. LSTMs for sequence related problems (predicting poses of humans using previous poses). Skilled in Python, TensorFlow, PyTorch, OpenCV, and designing AI-driven solutions for visual perception, generative modeling. Skilled in fine-tuning LLMs with hands-on experience using LangGraph for building multi- agent AI systems. Hands on experience with langgraph based agent workflows, tool calling and developing ReAct agents for troubleshooting PC related problems. EDUCATION Software Engineering School of Electrical Engineering and Computer Science , Islamabad , 3.24 (2022) INTERNSHIP EXPERIENCE Pakistan Telecommunication Company Limited (PTCL) 20-Jun-2025 - 08-Aug-2025 Redesigned PTCL’s Legacy Order Management System by developing a modern React frontend and refactoring the backend into a microservices architecture in .NET. Integrated complex database structures and stored procedures into backend services, ensuring accurate feature mapping and improved system performance. True Tech Solutions 01-Jun-2024 - 01-Sep-2024 Developed a Desktop app in Flutter to improve in-store order handling for eyewear retail stores. Implemented Clean Architecture, BLoC for state management, MVVM, and dependency injection, with Firebase for authentication and database management. Contributed to both frontend and backend of an cross platform app, ensuring scalability and maintainability FINAL YEAR PROJECT Troubleshooting Agent An agent which can solve the local environment,server or desktop related errors like dependency issues which effect productivity of a developer. The Agent can help in diagnosis and solution of any problem related to the computer or server, based on the query of the user. Hardware, OS and kernel aware solutions with automated command execution with user permission for problems faced during deployments on server, development environment setup, Containerization & Orchestration, Networking & Connectivity Issues, Performance & Monitoring Issues and Hardware-Aware Diagnostics Example capabilities include diagnosing and resolving Python, Node.js, CUDA, and system dependency conflicts; kernel module and driver mismatches; Docker and Kubernetes deployment failures; GPU, disk, and memory hardware issues; server misconfigurations (Nginx, Gunicorn, systemd); networking, firewall, and

Machine Learning/Computer Vision Engineer with a strong passion for AI, Large Language Models (LLMs), Experienced in developing keypoint detection models, image segmentation, and deep learning pipelines for real-world applications, training object detection models like fire and smoke detection, video analytics using opencv. Comprehensive understanding of 3D generation models (e.g., NeRFs), encoder architectures including Variational Autoencoders, Diffusion Models, GANs and data preprocessing techniques (PCA, LDA, UMAP) for dimensionality reduction and feature extraction. LSTMs for sequence related problems (predicting poses of humans using previous poses). Skilled in Python, TensorFlow, PyTorch, OpenCV, and designing AI-driven solutions for visual perception, generative modeling. Skilled in fine-tuning LLMs with hands-on experience using LangGraph for building multi- agent AI systems. Hands on experience with langgraph based agent workflows, tool calling and developing ReAct agents for troubleshooting PC related problems.