Zunaira Amir

Academic

Verbatim text

Zunaira Amir
+923301584805, zunairaamir777@gmail.com
Soan Garden, Islamabad.
L
~
Linkedln: www.linkedin.com[in[zunaira-amir-ba844223b
github.com/Zamehi
Education
Bachelor of Science (Computer Science)
!Major:
~enerative Al, NLP, Digital Image Processing, MLOPs, Blockchain, Machine Learning, Statistical
Modelling, Software Engineering, Database, DSA.
HAPC, Rawalpindi, Punjab
F.Sc (Physics, Chemistry, Biology)
I
'Projects
'Final Proiect: Student Assistant using NLP techniques processing Audio and video lectures[NLP, RAG, BERT,
[React, D3.JS]
Talk-Track-FYP, performs semantic search, summarization and Al-driven chatbot development, focusing on
retrieving relevant information from uploaded content.
lSemester Proiects:
DjangoTweet: A Twitter Clone:
~ Django-based social media platform replicating core Twitter functionalities.
NLP-Powered Data Similarity Engine:
~n Al-driven model to measure and compare text similarity using NLP techniques.
MedChat-RAG: Al Medical Chatbot :
~ RAG -based chatbot leveraging LLaMA for intelligent medical knowledge retrieval.
ISentimentScope: Movie Review Analysis:
~ text classification model predicting sentiments of movie reviews.
NeuralSpamShield: Spam Detection System:
lA. neural network-based solution to detect and filter spam messages.
1Sketch2Face: Conditional GAN for Face Reconstruction
lA. deep learning model using Conditional GANs to generate realistic faces from sketches, with a Flask GUI.
Work Experience
NLP Intern I IKNEX Lab, FAST Nuces, lsb. (Supervision of Dr. Amna Basharat)
Aug 2017 - Apr 2019
lA.chieved 80% text categorization accuracy by using Word2Vec, TF-IDF, Transformers, and
BERT for Arabic and English text similarity via N LP models.
'Skills & Tools
Professional Skills
Leadership, Interpersonal Skills, Communication.
rTechnical Skills
Python, Tensor Flow, Keras, BERT, Pytorch, OpenCV, RAG, LLM, SQL, Sci-kit Learn, SQL, Django
Rest Framework, Java including JDBC, Java, HTML, CSS, Socket programming, GO
'Achievements
Mentioned in Deans List for two semesters.
Tools
Git, Docker, Linux, WSL2, Kubernates
'Activities
President, FAST Gaming Club (24-25), Head Marketing, FQSS (23-24), Vice Head PR, IEEE (23-24)
Interests
K3aming
FAST NUCES ISLAMABAD CAMPUS
FAST SCHOOL OF
COMPUTING
FINAL YEAR PROJECTS
FAST NUCES ISLAMABAD CAMPUS
NewsNet: GraphRAG for Deep News Analysis
NewsNet is a GraphRAG project for deep news analysis that aims to map meaningful relationships
between news events from raw news content. Based on Natural Language Processing and
Knowledge Graph techniques, the system is designed to extract entities and map the complex
causal relationships to enable users to trace events through time, understanding their
development and impact. NewsNet addresses limitations in traditional news analysis by providing
event continuity and granular exploration of complex issues. The cutting-edge system utilizes
GraphRAG, a combination of graph databases and retrieval-augmented generation, to analyze
news data in depth. This approach allows for a deeper understanding of complex events, tracing
their history, and providing insightful context.
Features include:
- Deep News Analysis by extracting relevant entities and their related links from raw news data.
- Creating dynamic timelines for events, providing a historical perspective on news stories
- Graphical visualization of user search for better understanding the user's interest
Key Words: Natural Language Processing, Knowledge Graph, Retrieval-Augmented Generation, Event
Extraction, News Analysis
@
NEWSNET
UOIJl'W:MBBS.:
llSA•A...:DZW-9451
MUHAMMAD r&AJIAll AUIAS 1:W- M48
S 'E:D MUHAMMAI AU MUBMJll-Z566
DESIGN
5m'CRYls.MS:
DRAllSllADISLAM
MRM•lr£11MAll
i~i r---l--~--=====:=-l
HOW IT WORKS
) lm5t!LDll0901 I
L--~J-----
-~­
-
- ;\~
·;
\-
E•Tlll SUKU.PH
~1
.....
-
... ~
,..-
....
~
, I
TOOLS
* ~ ,
--. ~ ,
ml
ml
~m
Technology Used:
Python, React,Go,D3.JS, FIASS
Co-Supervisor Name:
Dr. Arshad Islam
Su ervisor Name:
Mr. Aqib Rehman
Group Members:
Mr. Usaid Ahmed (i121 - 0458}
Mr. Syed Muhammad Ale Mubar (i21- 2566}
Mr. Muhammad Fakhar Abbas (i21- 0448}
FAST NUCES ISLAMABAD CAMPUS
lnsuraSync
lnsuraSync is an Al-powered platform designed to streamline the insurance claims process,
particularly focused on car accident claims. The platform caters to two types of users: Customers
and Admins. The Customer View allows users to submit claims, track their progress, and receive
real-time updates through a user-friendly interface. The platform also provides automated repair
cost estimations, fraud and discrepancy detection, and a comprehensive self-service portal. If a
customer faces any issues, they can reach out to the admin team directly from the application.
The Admin View is tailored for insurance providers, enabling them to manage claims efficiently,
review flagged discrepancies, and access detailed analytics. Admins can also oversee the system's
Al-driven decisions and intervene when necessary to update or correct claim-related information.
Features include:
Automated Claims Processing: Reduces manual intervention in claim validation and approval.
Fraud and Discrepancy Detection: Uses Al to identify anomalies or potential fraudulent claims.
Repair Cost Estimation: Automatically calculates repair costs using accident data.
Self-Service Portal: Customers can submit claims, track updates, and manage their policy details.
Personalized Policy Recommendations: Offers customers tailored policy options based on their
profile and past interactions.
Admin Dashboard: Allows admins to manage claims, resolve flagged issues, and monitor system
performance.
Key Words:
Al-powered Claims Processing, Personalized Policy Recommendations, Fraud Detection, Automated Repair
Cost Estimation, Customer Self-Service Portal
Technology Used:
Flutter, Node.js, tensorflow, Firebase, Github,
Pyhton,Visual Studio
Supervisor Name:
Dr. Ali Zeeshan
Group Members:
Mehreen lsrar (i21 - 0594)
Aden Amar (i21 - 0755)
Ghulam Mustafa (i21 - 0832)
FAST NUCES ISLAMABAD CAMPUS
Mumtahin
Mumtahin is a project designed to automate the assessment process, reducing the manual
workload for both students and evaluators. It tackles challenges such as timeslot conflicts,
evaluator bias, and delays. The system utilizes facial recognition for student verification,
generates and evaluates answers, and monitors facial expressions to detect potential cheating.
Additionally, Mumtahin verifies submitted code, checks for plagiarism, and creates detailed
reports to ensure a fair and efficient evaluation process.
Features include:
•Facial Recognition: Verifies student identity during the demo to ensure the person presenting is the
correctly identified student.
• Plagiarism Detection: Analyzed the assignment for plagiarism.
•Slot Management: Automates the scheduling and rescheduling of demos based on student and evaluator
availability.
• Submission Checking: Verifies that the submitted code matches the assignment requirements before
evaluation.
•Generate Report: Produces detailed reports with grades, answers, facial expression analysis, and code
submission.
•Video demo: The student will join the meeting at the allocated time and complete the quiz.
~--·••1
C T F A M
M
F MAFRS
C
S UPERVISOR
AA.fl~~l ik Al", b A'°"i
(«'1 1 - ~b..,t_>)
Affan A n w a r (21.1 254 8)
M v h1:1rr1rf1e:td H<:t trHf1 i:H.:I (? l l -7~ 815 )
Mr. S aad Sal rnan
New E ra of Eva l uation!-.!
A l- t l<i""''' ;., . ~ ;.t::, , ,. , .. ,., '"'""'";• Ii<"' vviLh , .,.,1- 1 ;,, ,.., ,J.,,,.,,,., q u ;,..,.. .. ,., .. , ,d p l"t:;;.,;,,,"
,J..,,.,,,,;,,,,_ ,., .. .., ... ,.1;,,;,,.t:,..••;••l . ..,, ...... ,, 1,,.,i.;,,, ...
ARCHITECTUR E
~ ~
{ .- =-;;:=';'.~: I
:;i::E· I ,:.~_,, :~~:.:;;_·
~ ~
.._,
. -
!\ ~ ~ ~
-
,,..
<S>
<T>
... , .......
a
~'A"~" ~'
""
;::::
i
~
~
· ~ ·-'"•'-"'"~
:m
i.,'$.~'Y
c TOOLSANDTECHNOLOGY
Technology Used:
Git ,Pytorch ,Python,MongoDb,Flask,
Node js,react
Supervisor Name:
Mr. Saad Salman
Group Members:
Aaimlik Abbasi (i21 - 2540)
Muhammad Hammad (i21-25

AI enrichment