Zain Ul Abdeen
> INIT_PROTOCOL: AI Systems Engineer...
CGPA: 3.65
STUDENT @ GIKI
AI-powered textile pattern generation suite using VAE, GANs (DC, WGAN-GP, cGAN), Latent DiT, and Flux API with a React frontend dashboard.
A specialized generative suite for fashion designers featuring 6 locally-trained models and Flux API integration. Includes a complete data pipeline (DTD, 5,640 images, 47 classes), quantitative evaluation (FID, IS, SSIM), containerized REST API, and CI/CD via GitHub Actions.
AnyLogic-based IoT network simulation with embedded ML for latency prediction, anomaly detection, forecasting, and adaptive Q-learning load management.
A 5-layer IoT simulation platform: MQTT-style packet flows, transpiled Java ML models (Random Forest, One-Class SVM, BiLSTM) for inline inference, tabular Q-learning for adaptive congestion control, FedAvg federated learning across 3 edge nodes, and a dual-surface NOC + React web dashboard. Uses RT-IoT2022 dataset (4,132 flows) with GAN-style DDoS stress injection.
Financial ML platform covering 26 models across 8 tickers. Regression, BiLSTM forecasting, KMeans, and SMOTE-based fraud detection.
End-to-end financial machine learning platform with 26 trained models across 8 stock tickers. Covers regression analysis, BiLSTM time-series forecasting, KMeans clustering, and SMOTE-based credit card fraud detection with containerized deployment.
Agentic Research Assistant taking hypothesis, running ArXiv/Semantic Scholar QC, then generating operationally-grounded experiment plans.
An agentic research assistant that accepts a hypothesis, runs quality-controlled searches via ArXiv and Semantic Scholar APIs, performs literature synthesis, and generates operationally-grounded experiment plans with full citation support.
Pipeline combining CLIP (embeddings), FAISS (search), Neo4j (graph), LangChain ReAct agent, and BLIP-2/LLaVA (visual QA).
A multimodal retrieval-augmented generation pipeline combining CLIP embeddings, FAISS vector search, Neo4j knowledge graphs, LangChain ReAct agents, and BLIP-2/LLaVA for visual question answering over complex document collections.
Automated deployment of DOMjudge using Bash and Docker Compose. CLI-driven control & judgehost scaling.
A fully automated deployment solution for the DOMjudge competitive programming judge system. Features CLI-driven control, dynamic judgehost scaling, and one-command setup with Docker Compose orchestration.
GitHub Actions logic connecting React, Node.js, and DB to AWS EC2 via ECR (Staging/Testing envs).
End-to-end CI/CD pipeline using GitHub Actions to build, test, and deploy a full-stack app (React + Node.js + DB) to AWS EC2 via ECR. Supports isolated staging and testing environments with automated rollback.
An AI-powered dining revolution blending smart automation with seamless restaurant management.
An AI-powered smart restaurant management system featuring automated ordering, real-time kitchen management, intelligent menu recommendations, and seamless payment integration. Built with Flask backend and MongoDB for scalable data persistence.
EduMate is an AI-driven web app that helps students study smarter by summarizing notes & planning.
An AI-driven educational web application powered by Google Gemini. Helps students study smarter through automated note summarization, intelligent study planning, flashcard generation, and personalized learning paths.
Standard banking operations via a C++ OOP architecture.
A comprehensive banking system implementing core operations — account creation, deposits, withdrawals, transfers, and statement generation — using clean C++ OOP architecture with inheritance, polymorphism, and file-based persistence.
C++ console-based 2D race car game with lane logic and custom algorithms.
A console-based 2D racing game built in C++ featuring real-time lane-switching mechanics, obstacle avoidance algorithms, dynamic difficulty scaling, and score tracking — all implemented using custom data structures and algorithms.
Flask-based deep learning web app that recognizes handwritten characters (CNN via EMNIST).
A deep learning web application using a Convolutional Neural Network trained on the EMNIST dataset. Users draw characters on a canvas and get real-time predictions. Deployed as a Flask web app with interactive frontend.
CGPA: 3.65
Relevant Coursework: Machine Learning, Deep Learning, Data Structures & Algorithms, Database Systems, Computer Vision, Big Data Analytics.
Worked remotely on building scalable machine learning solutions, integrating data pipelines, and implementing AI agents.
Developed core AI functionalities, optimized deep learning models, and handled data preprocessing tasks remotely.
Developed and fine-tuned NLP pipelines for text classification, entity recognition, and semantic similarity. Worked on transformer-based models and contributed to production-ready language processing modules.
Stanford University | 2023
3-course specialization by Andrew Ng covering supervised learning (linear/logistic regression, neural networks), advanced algorithms (decision trees, ensemble methods), and unsupervised learning (clustering, anomaly detection, recommender systems).
DeepLearning.AI | 2023
Foundational mathematics for machine learning covering linear algebra (vectors, matrices, eigendecomposition), multivariate calculus (gradients, backpropagation), and probability & statistics (Bayes' theorem, distributions, PCA).
DeepLearning.AI | 2023
Professional certificate covering TensorFlow for building CNNs for computer vision, NLP with RNNs/LSTMs and word embeddings, time-series forecasting, and deploying production-grade deep learning models with TF Serving and TFLite.
DeepLearning.AI | 2023
Professional certificate on leveraging LLMs and generative AI for software development — prompt engineering, AI-assisted code generation, debugging workflows, automated testing, and integrating AI copilots into development pipelines.
Google | 2023
Google's course on data analysis using R programming — data cleaning with tidyverse, exploratory data analysis, statistical visualization with ggplot2, R Markdown reporting, and working with structured datasets for business insights.
Microsoft | 2023
Certified proficiency in the Microsoft Office suite — advanced Excel (pivot tables, VLOOKUP, macros), Word document formatting, PowerPoint presentation design, and data management with Access and Outlook productivity workflows.
Oracle | Associate
Oracle certification covering AI and ML fundamentals — deep learning, NLP, computer vision, Oracle Cloud Infrastructure (OCI) AI services, generative AI concepts, responsible AI practices, and deploying ML models on OCI.
IBM
IBM course on RAG architectures — vector databases, embedding models, document chunking strategies, retrieval pipelines, prompt augmentation with context injection, and building grounded LLM applications that minimize hallucinations.
Google | 8 Courses
Google's 8-course professional certificate covering the full data analytics lifecycle — Python, exploratory data analysis, statistical modeling, regression analysis, machine learning with scikit-learn, and Tableau visualization for executive dashboards.
Coursera | Professional Certificate
Professional certificate covering fundamentals of PyTorch and neural networks, advanced architectures, and model deployment techniques, using real-world AI systems for CV and NLP with TorchVision and Hugging Face.
Coursera | DeepLearning.AI
5-course specialization covering Neural Networks, Hyperparameter Tuning, Structuring ML Projects, CNNs, and Sequence Models (RNNs, LSTMs, Transformers). Practical applications in Python and TensorFlow.
I'm currently looking for new opportunities in AI engineering and MLOps. My inbox is open.