← Back to cohort

Mudassar Manzoor

FAST · 2025

writetomudassarawan@gmail.com

Phone

+923165511188

—

GitHub

—

Academic

Program

—

CGPA

—

Year

2025

Education

—

Address

—

DOB

—

Verbatim text

The exact text the LLM saw on the page (or the booklet text from the old import). This is what powers semantic search.

r
Mudassar Manzoor
'
+923165511188, writetomudassarawan@gmail.com
G-11/3, Islamabad
Linkedln: htt~s:L[www.linkedin.com[in[mudassar-alblcldll
Education
Bachelor of Artificial Intelligence (Al)
Major:
Generative Al, Computer Vision, Knowledge Representation & Reasoning, Artificial
Intelligence, Database Systems, Fundamentals of Software Engineering, Parallel Distributed
Computing, Computer Networks, Reinforcement Learning, Machine Learning
APSACS Muzaffarabad
F.Sc (Physics, Computer Science, Mathematics)
APSACS Muzaffarabad
Matriculation (Physics, Chemistry, Computer Science)
Projects
-;nal Proiect: DashGrab [kalman filter, yolovllx, open pose, detectron 2)
DashGrab is a cashier-less store system for clothing brands that uses CCTV and computer vision to track items
picked up by customers in real time, automating billing and inventory like Amazon Go.
Proiects:
Whatsapp Bot builder:
Made a bot builder where users come with their shop info ( inventry, shop information (what they do) ), and my
bot builder make a bot for them specific for their phone number, connects it, and they are good to go. Now their
users can message their bot and have a conversation.
Multimodal Sentiment Analysis:
Implemented a Neural Network from scratch using pytorch for sentiment analysis using textual and visual data.
Evaluation: Achieved an accuracy of 94.7% on the test set,
outperforming unimodal baselines.
Deepfake Detection :
I Implemented a Convolutional Neural Network (CNN) to detect deepfakes in a video dataset.Dataset: Used the
DFDC dataset containing 2,000 video samples, with 1500 real videos and 500 deepfake videos. Approach:
Preprocessed videos to extract frames, then trained a CNN model with 4 convolutional layers and 3 fully connected
layers.
Results: Achieved an accuracy of 89% on the test set.
CycleGAN Implementation for Person Face Sketches:
Conversting face images to sketches and sketches back to face images using cycle gans
Experience
Associate Machine Learning Engineer at I MARAT
I August 2024- present
Skills & Tools
Professional Skills
rreamwork & Collaboration, Interpersonal Skills, Communication.
Technical Skills
Python, C++, Assembly Language, C, SQL, Bootstrap and HTML & CSS, Git,
MySQL, Visual Studio Code, PyCharm, SQLite, Protege, GraphDB, VS Code,
Uupyter
Notebook, TensorFlow, & Pytorch, NumPy, Pandas, Scikit Learn, OWLReady2,
Keras
'nterest
chess, conspiracies, islam, philosophy
FAST NUCES ISLAMABAD CAMPUS

AI enrichment

Mudassar Manzoor is an Associate Machine Learning Engineer with a Bachelor's degree in Artificial Intelligence and experience at I MARAT. He specializes in computer vision, deep learning, and generative AI, demonstrated through projects involving object detection, sentiment analysis, and deepfake detection.

Skills (AI)

["Python", "PyTorch", "TensorFlow", "Computer Vision", "Machine Learning", "Deep Learning", "C++", "SQL", "Git", "Generative AI", "Reinforcement Learning", "Object Detection", "Sentiment Analysis", "Deepfake Detection", "CycleGAN", "YOLO", "OpenPose", "Detectron 2"]

Status: ai_done

Provenance

Source file: —
Created: 1777724106