Manahil Ahmad

Academic

Career

Verbatim text

inference. Tools/Skills: 
Python, 
CUDA, 
PyTorch, 
Keras, 
Deep 
Learning, 
GPU 
Acceleration, 
Image 
Processing.
Certificate:https://drive.google.com/file/d/1-wgvw7zO5cLq_6zO8dclqCib4k07IBBS/view?usp=sharing
Github:https://github.com/manahild/MNIST-CLASSIFCATION-VIA-MLP-IN-CUDA
Smart Agritech Lab, SINES — Deep Learning and Computer Vision Intern
18-Jul-2024 - 12-Sep-2024
Developed a computer vision system to identify Holstein Friesian cattle by coat patterns. Managed end-to-end ML pipeline: data
collection, annotation with LabelMe, data augmentation, and model training. Trained YOLOv8 on a custom dataset, achieving high
accuracy in cattle identification, enabling automated livestock monitoring. Tools/Skills: Python, YOLOv8, Computer Vision, Image
Processing, Deep Learning, LabelMe. Certificate: https://drive.google.com/file/d/15VWnxKa1aFzKbd1zQs8OBLOBrFvT48mB/view?
usp=sharing Github:https: //github.com/manahild/Yolov8_for_Cattle_Identification_via_Coat_Pattern-
Team Alif – Formula Student NUST — Electrical Team Member
19-Oct-2023 - 12-Jun-2024
Contributed to the design and development of an electric racing car, applying electrical engineering principles for vehicle power
systems. Analyzed and optimized battery performance, designed wiring layouts, and supported integration of sensors and controllers
for reliable system operation. Coordinated with multidisciplinary teams to ensure timely project execution and alignment with
competition standards. Tools/Skills: MATLAB, Electrical Design, Battery Optimization, Embedded Systems, Team Coordination.
Millennium Campus Network (Partnered with UNAI) — Millennium Fellow
21-Aug-2024 - 01-Jan-2025
Led the “Restoring Schools, Inspiring Minds” project aligned with UN SDG-4, providing educational support to underprivileged
students. Coordinated community outreach, awareness sessions, and educational activities, improving access to learning resources.
Developed project management, leadership, and team coordination skills while achieving measurable social impact. Skills:
Leadership, 
Project 
Management, 
Community 
Engagement, 
Strategic 
Planning. 
Certificate:
https://drive.google.com/file/d/16PUO0OuN87GRfzuqGkuhv1rlV0hFGaWz/view?usp=sharing 
Links:
https://www.millenniumfellows.org/fellow/2024/nust-pk/manahil-ahmad
FINAL YEAR PROJECT
"Intelligent Control of Stacked Intelligent Metasurfaces-Assisted Wireless Networks using Deep
Reinforcement Learning"
Designed and implemented a SIM-enabled Integrated Sensing and Communication (ISAC) system model, enabling simultaneous
communication and sensing within a unified wireless framework for 5G/6G next-generation networks. Formulated point-target and
extended-target sensing performance optimization using Cramér–Rao Bound (CRB) alongside communication throughput, balancing
high data rates with reliable connectivity in dense network scenarios. Developed a Deep Reinforcement Learning (DRL) optimization
framework to learn optimal phase shift configurations of stacked intelligent metasurfaces, improving signal quality, spectral efficiency,
and network reliability under realistic power, noise, and channel constraints. Modeled complex wireless propagation environments,
simulating both communication and sensing channels, to evaluate trade-offs between sensing accuracy and communication
performance critical for 5G/6G deployment. Implemented and trained DRL agents in Python and PyTorch, focusing on reward design,
policy learning, and convergence behavior for real-world ISAC optimization. Conducted extensive system-level simulations to
analyze throughput, latency, reliability, and sensing accuracy across multiple network configurations. Research outcomes contributed
to a peer-reviewed publication accepted at IEEE GLOBECOM 2025, titled: “DRL-Based Phase Shift Optimization for SIM-Enabled
Wireless Systems”, demonstrating applicability to telecom industry challenges such as enhancing data rates, coverage, and reliability
in next-generation networks.
TECHNICAL EXPERTISE
Artificial Intelligence, Machine Learning & Deep Learning
Complete Coursera Specialization in Reinforcement Learning & Deep Reinforcement Learning Strong command of Machine Learning
and Deep Learning theory and practice Supervised Learning, Deep Neural Networks, CNNs, RNNs, Transformers Model training,
evaluation, hyperparameter tuning, and performance opti ...
Computer Vision & Image Processing
Classical Computer Vision: filtering, edge detection, feature extraction Deep Learning–based Vision: CNNs, Vision Transformers
(ViT) Object Detection, Image Classification, Custom Dataset Training Tools & Models: OpenCV, YOLOv8, YOLOS Image processing
in Python and MATLAB

AI enrichment