Computer Science • Software • ML

Hi, I’m Sarah.

An aspiring computer science graduate student who loves building software solutions and solving challenging problems.

View projects Get in touch

Email [email protected]

Location Illinois, US

GitHub LinkedIn

Graduate Student

Full-stack • Data Science • ML • Research

Featured Projects

Research Project

∞

Curiosity

Open to opportunities Internships / Research

Focus Computer Eng + AI/ML

Machine Learning • Full Stack Development • Data Science • Computer Vision • NLP • Cybersecurity • Machine Learning • Full Stack Development • Data Science • Computer Vision • NLP • Cybersecurity •

About

Objective & what I’m looking for.

Objective

Aspiring computer science graduate student with a passion for developing innovative software solutions and solving challenging problems. Seeking an opportunity to apply strong programming skills, analytical thinking, and technical knowledge while continuously learning and collaborating on impactful projects.

Highlights

Data Scientist Intern experience (Python, SQL, ML, reporting)
Research: Diabetic Retinopathy Classification using DL and handcrafted algorithms
Awards: RTA Hackathon (2nd place), Deputy Principal’s Award in Undergrad

Moments

Hackathons

Education

Degrees and academic awards.

University of Illinois at Chicago

Master's in Computer Science — Aug 2025 to Jun 2027 (pursuing)

Heriot-Watt University, Dubai

BSc. Computer Science (Hons) — Sep 2021 to Jun 2025

Work Experience

Internship and responsibilities.

Intellipaat Software Solutions Pvt. Ltd.

Jul 2022 – Jul 2023

Data Scientist Intern

Analyzed large datasets to identify trends and insights supporting data-driven decisions.
Developed and optimized machine learning models to improve predictive accuracy.
Cleaned, processed, and visualized data using Python, SQL, and visualization libraries to generate actionable reports.
Collaborated with cross-functional teams to deliver analytical solutions that streamlined business operations.

Research

Undergraduate research project.

Fusion of Deep Learning and Handcrafted Features for Diabetic Retinopathy Classification

2024 – 2025

Heriot-Watt University, Dubai

Built a hybrid classification framework combining CNN features with handcrafted image features to improve detection accuracy.
Collected, cleaned, and augmented retinal fundus datasets; addressed class imbalance.
Extracted GLCM, color histograms, blood vessel patterns, and fused them with deep embeddings.
Fine-tuned CNNs with transfer learning and dropout; documented experiments and results in a technical report.

Diabetic Retinopathy Artificial Intelligence Computer vision

Projects

Selected builds and roles.

Multi‑Task Transformer Fine‑Tuning

Code

Deep Learning / NLP Engineer — 2025

Fine‑tuned DeBERTa‑v3 for multi‑task emotion regression & polarity classification.
Designed custom multi-head architecture; improved classification accuracy to 86.7% (+4.1%).
Built custom Trainer pipeline with GPU acceleration, mixed precision, and Optuna tuning.

Transformers Multi‑Task Learning HuggingFace Optuna

Corpus-Based Empathetic Chatbot

Code

NLP Engineer — 2025

Built an emotion-aware chatbot using WASSA 2024 dataset and SentenceTransformers.
Developed weighted retrieval combining semantic, emotional, and polarity similarity.
Achieved strong semantic similarity (BERTScore-F1 ≈ 0.85) in dialogue generation.

NLP Retrieval-Based Sentence Transformers Emotion-Aware

Skin Lesion Classification (HAM10000)

Code

ML & Computer Vision Engineer — 2025

Classified 7 lesion types (HAM10000) comparing Classical ML (SVM/RF) vs. Deep Learning.
Deep Learning (MLP) achieved ~95% accuracy, significantly outperforming Random Forest (~71%).
Performed extensive EDA and unsupervised learning (K-Means, DBSCAN) on class imbalance.

Computer Vision Deep Learning CNN MLP Medical Imaging

Hydra Brute-Force Attack Simulation

Code

Cybersecurity Analyst — 2025

Conducted controlled brute-force attacks on MySQL, SSH, and WordPress using Hydra.
Demonstrated vulnerabilities in weak credentials and misconfigured services in a Kali Linux lab.
Implemented hardening: SSH key-based auth, MFA, Fail2Ban, and firewall policies.

Penetration Testing Ethical Hacking Hydra Network Security

Customer Churn Prediction System

Code

Machine Learning Engineer — 2024

Built an end‑to‑end churn prediction model for telecom data using XGBoost.
Handled class imbalance and performed rigorous feature engineering.
Optimized via ROC‑AUC evaluation, improving recall while minimizing false positives.

Supervised Learning Classification Imbalanced Data ROC‑AUC XGBoost

Deep Image Classification with Transfer Learning

Code

Deep Learning Engineer — 2025

Designed and fine‑tuned pretrained ResNet18 and ResNet34 architectures on CIFAR‑10 using GPU acceleration.
Implemented data augmentation, learning rate scheduling, and full network fine‑tuning to improve accuracy from 78% to 85%+.
Analyzed per-class accuracy, confusion matrices, and model performance trade-offs between shallow and deep architectures.

Deep Learning CNN Transfer Learning PyTorch Computer Vision

Fake News Detection System

Code

Machine Learning Engineer — 2026

Built a high-accuracy fake news classifier using TF‑IDF with n‑grams and Linear SVM on a 44K+ article dataset.
Performed advanced preprocessing, stratified splitting, confusion matrix analysis, and model evaluation using precision, recall, and F1-score.
Achieved 99%+ accuracy while analyzing dataset bias and real-world generalization implications.

NLP Linear SVM Text Vectorization Model Evaluation

Hybrid Intrusion Detection System

Code

Machine Learning Engineer — 2026

Built a two‑stage IDS on 2.8M+ network records: binary (Benign vs Attack) then multi‑class attack classification using Random Forest & Logistic Regression.
Random Forest achieved 99.88% accuracy and ROC‑AUC of 0.9988 with near‑zero false positives — production‑grade performance.
Handled severe class imbalance, preserved rare attack types (Heartbleed, SQL Injection, Infiltration).

Cybersecurity Machine Learning Data Analytics

Hybrid Movie Recommendation System

Code

Machine Learning Engineer — 2025

Developed a hybrid recommendation engine combining collaborative filtering (SVD) and content-based filtering (cosine similarity).
Implemented matrix factorization, genre-based similarity modeling, score normalization, and weighted blending.
Compared baseline vs hybrid performance and deployed model persistence for scalable inference.

Recommender Systems Matrix Factorization Feature Engineering

Resume Screening & Candidate Ranking System

Code

Machine Learning Engineer — 2025

Built an end‑to‑end NLP system to automatically classify and rank resumes using TF‑IDF, cosine similarity, and supervised learning.
Implemented resume category prediction, similarity scoring, and job‑matching functionality.
Designed preprocessing pipelines, model evaluation workflows, and persistent model storage for deployment readiness.

NLP TF‑IDF Text Classification Cosine Similarity

Collabrain

Code

Full-Stack Developer — 2023–2024

Built a web + mobile app for social networking and collaborative brainstorming.
Integrated real-time interaction, teleconferencing, and office-suite capabilities with incentives.
Contributed to requirements, UML modeling, frontend development, system design, and project costing.

Full-stack UML System design

Pets&Me

Case study

UI/UX Designer — 2023

Designed a user-centric app with pet profiles, task notifications, GPS, and health tracking.
Enabled real-time monitoring via wearable devices / phone sensors for proactive health insights.
Iterated UI/UX with feedback to improve engagement and usability.

UI/UX Figma Mobile

Salem

Code

Mobile App Developer — 2022

Built a safety-focused app using sensors, GPS, AI/ML to monitor delivery personnel health & driving behavior.
Provided instant feedback, improvement tips, and rewards for safe driving adherence.
Optimized performance in VS Code to reduce lag and improve UX.

Mobile Sensors + GPS Optimization