Ahsan Taimoor Ghazi

Academic

Career

Verbatim text

Ahsan Taimoor Ghazi
Cell: 03326448559 |  Email: dev.ahsanghazi@gmail.com
LinkedIn: https://www.linkedin.com/in/devahsanghazi/
Address: Nust Hostels H-12 Islamabad , Islamabad , Pakistan
PROFESSIONAL PROFILE
Agentic AI Engineer with 6 months of hands-on experience building AI systems, agent workflows, and computer vision models. I
enjoy working on practical problems and creating systems that feel smooth, clean, and useful. I have completed MLOps practices,
enabling me to deploy, automate, and manage end-to-end AI solutions effectively.
EDUCATION
Bachelor of Software Engineering
Seecs , Islamabad , 2.9 (2022)
INTERNSHIP EXPERIENCE
Conceptuals Academy
03-Nov-2025 - 19-Dec-2025
Developed an AI-powered chatbot for an academic organization offering O-Level and A-Level education, integrated with the existing
Learning Management System (LMS). Designed the chatbot using a retrieval-augmented generation (RAG) approach to answer
academic and course-related queries by grounding responses in institutional content. Implemented agent-based logic to manage
query handling, information retrieval, and response generation workflows. Deployed the system using MLOps practices, focusing on
reliable model serving, automation, and maintainability in a real-world educational environment.
FINAL YEAR PROJECT
Emotion Detection Based on Micro Expression
Micro-Expression Emotion Recognition using Vision Transformers Designing an emotion detection system based on facial micro-
expressions using a Vision Transformer (MaeFace) architecture. Pretraining the model using self-supervised learning on macro-
expression datasets to learn robust facial representations. Fine-tuning the pretrained model on micro-expression datasets including
CASME II, CASME III, SMIC, and SAMM. Implementing knowledge distillation to transfer learned representations from a large
MaeFace (teacher) model to a lightweight MobileViT (student) model. Optimizing the student model for real-time, low-latency
inference on resource-constrained devices.
TECHNICAL EXPERTISE
Agentic AI
Design and develop agent-based AI systems capable of planning, tool usage, and multi-step reasoning. Build structured agent
workflows to solve real-world tasks using modular, extensible architectures.
Retrieval-Augmented Generation (RAG)
Develop RAG pipelines for grounding LLM responses using external knowledge sources. Implement document ingestion, embedding
generation, vector search, and response synthesis to improve accuracy and relevance.
Computer Vision
Build computer vision models for image and video analysis using deep learning techniques. Work with CNNs and Vision
Transformers for tasks such as facial analysis, micro-expression recognition, and real-time inference.
Model Optimization & Deployment
Build computer vision models for image and video analysis using deep learning techniques. Work with CNNs and Vision

AI enrichment