Ushna Nadeem

Academic

Verbatim text

Ushna Nadeem
+923165168310, ushna.2002.ushna@gmail.com
-:-
DHA Phase 7, Islamabad.
Linked In: https://www.linkedin.com/in/ushna-nadeem-Slb838283/
Education
Bachelor of Science (Software Engineering)
Major:
~ecure Software Development, Software Programming, Software Quality Assurance, Software
V\rchitecture,
Artificial
Intelligence,
Human-Computer
Interaction,
Software
Project
Management & Data-Driven Software Engineering.
PakTurk Maarif Int. Schools and Colleges, Islamabad
F.Sc. (Physics, Chemistry, Mathematics)
Islam abad Model College for Girls, Islamabad
Matriculation (Physics, Chemistry, Biology)
Projects
.. inal Proiect: PhysiQuest VR [Unity, MetaQuest, Blender, Oculus, MySQL, Adobe Illustrator]:
A project for gamifying physics learning t hrough interactive VR experiences, to creat e a more engaging, hands-on and
intuitive way for students to grasp complex physics principles, making learning not only more effective but also fun.
Semester Proiects:
PawPrint Al - Dog Breed Recognition System [Python]:
Developed an Al project utilizing supervised learning techniques with ResNet-50 and VGG16 models to recognize dog
breeds accurately.
Feature Selection - Genetic Algorithm [Python]:
Developed a Genetic Algorithm-based feature selection approach to optimize neural network performance on
emotion classification using the RAVDESS dataset.
MedicMaster- Hospital Management System [Node.js, React]:
Developed the patient module of a hospital management system using Node.js and React, enabling appointment
:.cheduling, prescription access, and patient management.
RedDrop - Blood Bank Management System [Java, JavaFX, Figma]:
Designed a sophisticated system using Java, JavaFX, and Figma, ensuring efficient blood bank operations and donor
management.
Work Experience
Ul/UX & Graphic Design Intern, iENGINEERING, Islamabad.
Jul 2024 - Aug 2024
Working as a UI intern, I developed high-fidelity mockups, gained skills in Adobe Illustrator
and Figma, and contributed to a website redesign, boosting engagement by 15%.
MLOPs Teacher Assistant, FAST University, Islamabad.
Feb 2025 - May 2025
Assisted in the Machine Learning Operations course, guiding student s on model
deployment, Cl/ CD pipelines, and cloud-based ML workflows.
Skills & Tools
Professional Skills
Leadership, Int erpersonal Skills, Communication.
Technical Skills
C++, JavaScript, Python, HTML, CSS, JavaFX, Node.js, Express.js, and React, Docker, Kubernetes,
Cl/CD pipelines, DVC, Jenkins, Version control (Git), Agile methodologies.
Achievements
Awarded with a 25% scholarship of tuition fee at college level.
Training I Certification
2021: "Create Charts and Dashboard using Google Sheets" by Coursera Project Network through Coursera.
2015: "National English Contest" by GSEA (Global System for Educational Assessment) at Troff Level.
Activities
Member for 'Security' in NaSCOn'22.
Interests
Playing strategy and problem-solving games.
FAST NUCES ISLAMABAD CAMPUS
FAST SCHOOL OF
COMPUTING
FINAL YEAR PROJECTS
FAST NUCES ISLAMABAD CAMPUS
Virtulectra
Virtulectra is a web-based application geared towards teachers who want to lessen their burden
and students who would greatly benefit from interactive lectures. The platform addresses the
problem of a lack of interactive and efficient teaching tools, which leads to increased workload for
teachers and low student engagement.
The application supports two distinct user views: Teacher and Student. Teachers can manage
classrooms, upload course materials (PDF, DOCX), generate lectures, create customized quizzes,
and monitor student engagement. Students can join classrooms using unique codes, attend
Al-delivered lectures, ask questions in real-time, respond to pop-up questions, and complete
automatically graded quizzes.
Features include:
- Real-Time Interaction Module: Enables natural, interactive voice conversations between the
system and students during live lectures, with real-time question-answering
- Automated Content Creation Module: Generates comprehensive lecture scripts, notes, and
customized quizzes based on uploaded course materials
- Student Monitoring Module: Tracks student engagement using scheduled pop-up questions
during lectures
- Web Portal Module: Provides teacher dashboard and student interface with easy navigation to all
platform features
Key Words: Al-Powered Education, GraphRAG, LLM, Speech-to-Speech, Student Engagement,
Microservices Architecture
~
Virtulectra
Empowe<ing t eacher$by5implifyingWOf\.:loadar>d
~
""""''"'"''" '""b"'"""'""'°"'""~'o",.''' [
,.,
more produc:uve te«:h109 erwironment, ult imately luch119
_!{_ir
to improved educat!Oflal rE$.lf\s,
J
Service-Oriented Architecture
·-- *"-- ·--
d-
Application Workflow
'"'"""' ..._ .. _
Tools and Technologies
-
Project Timeline
··-·--
··~ ·-
Technology Used:
PyTorch, MERN, Python,
Neo4j, Ollama, LangChain
Su ervisor Name:
Dr. Shahela Saif
Grou Members:
Moiz Akhtar {i21- 1104)
Musa Haroon {i21- 1206)
Muhammad Abdullah {i21- 1215)
FAST NUCES ISLAMABAD CAMPUS
lnsightWire
InsightWire is a comprehensive news aggregation and analysis platform designed to improve how
users consume and understand political news. It addresses issues like bias in news reporting,
fragmented information, and lack of contextual background by offering news aggregation, bias
detection, story comparison, and contextual analytics.
Key Features
1. News Aggregation & Personalization
o
Collects and categorizes news from multiple sources.
o
Allows users to personalize their news feed based on preferences.
o
Highlights locally relevant political news.
2. Chatbot & User Interaction
o
AI-powered chatbot that summarizes articles and answers user queries.
o
Users can provide feedback and rate articles for content improvement.
3. Bias Detection & Story Comparison
o
Identifies political bias (left, center, right) in news articles.
o
Enables users to compare different perspectives on the same story.
o
Includes a voice-over feature for improved accessibility.
4. Contextual Background & Analytics
o
Generates timelines to show historical background for news events.
o
Provides analytical insights to detect trends and forward-looking predictions in news
articles.
lns ig ht\/Vire aims co c re ate a c ompreh10>nsi v e news p l alforrn lllal
addresses issu es of bias and lack o f c ontext. It a ggreg a tes n e w s from
g lobal s o urc es, hiChlig h l"> bias i:-s, provides c ontext and makes data
d riven an.:1lys i.s
~
WcbScr.,pin~
C .. tCl!Orization & T ar;r.:inr:
+---
Dio.pl.~y
• L LM a n d RAG
• Cailt.·e.ori.wlion
ln l c ;>.n'llion
• S torv
C ompari:;.on
• Loc a l N e ws
Hi,i:hlig:h t ing:
Backgroun d
• N t.:w ,; A n ,.lylic.;:.;
.. 0 * 0 PyTorch
n • dc ®
ak~;A ~
spaCy ..._.... ~ ~'
Technology Used:
Python, Node.js, Express.js, React.js,
MongoDB, RAG, Scrapy, Spacy, Pytorch,
LangGraph, LangChain
Supervisor Name:
Ms. Amna lrum & Ms. Kainat Iqbal
Group Members:
Muhammad Umar lhsan (211-1113)
Muhammad Sameer (211-1185)
Junaid Jamshaid (211-1203)
FAST NUCES ISLAMABAD CAMPUS
TracknRetrieve
TracknRetrieve is a Lost and Found app designed to assist users in reporting and
recovering lost items efficiently. The app features two views - User View and Admin View.
The former allows users to report lost or found items by providing descriptions and
images. The system uses Al-powered algorithms to match lost items with found ones
based on pattern recognition and similarity analysis. Users receive notifications when a
potential match is found. If the system fails to provide an accurate match, users can
escalate their query to the admin for manual review. The latter view is for the admin, who
verifies reported items, resolves disputes, and updates the system database to enhance
future matching accuracy.
Features include:
•
Al-driven matching for lost and found items based on descriptions and images.
•
Notifications for potential matches.
•
Location-based filtering for lost and found reports.
•
Admin moderation to verify and manage reported items.
•
Continuous database upda

AI enrichment