Abdullah Khaliq

Computer Science Student & Full-Stack Developer Specializing in AI Workflows and Distributed Systems

Based in: Lahore, Pakistan

A Computer Science student at LUMS with a strong foundation in software engineering, machine learning, and distributed hash tables. He has successfully delivered full-stack attendance systems and automated RAG pipelines using modern tools like Flutter, Node.js, and Databricks.

Skills

Technical

Workflow Automation & Pipelines

GitHub

VS Code

n8n

Python

TypeScript

RAG

Node.js

Postman

Java

JavaScript

HTML/CSS

MySQL

Machine Learning

Firebase

Software Engineering

NumPy

React

Scikit-learn

C/C++/C#

Android Studio

Databricks

Quantum Computing

Projects

Peer Tutoring Marketplace

Developeracademic

Built an Android app in Java that connects students with TAs. A student posts a request for a subject and gets matched with an available TA, and Firebase Realtime Database keeps the listings and matches in sync across devices. Ran the build as Agile Scrum sprints, with backlog grooming and iterative releases, and shipped a working app by the semester deadline.

AndroidJavaFirebaseFirebase Realtime DatabaseAgile Scrum

Feed-Forward Neural Network

Developeracademic

Wrote a two-layer feed-forward network from scratch in NumPy, coding the forward pass, ReLU and softmax, cross-entropy loss, and mini-batch backpropagation by hand. It reached over 75% test accuracy on Fashion-MNIST. Rebuilt the same network in PyTorch with nn.Module, Adam, and DataLoader, then tuned the learning rate, hidden-layer size, batch size, and optimiser (Adam, SGD, RMSprop) to push test accuracy past 85%.

PythonNumPyPyTorchnn.ModuleAdamDataLoaderSGDRMSprop

Peer-to-Peer File Sharing System

Developeracademic

Built a ring-based distributed hash table in Python that uses consistent hashing to spread key-value storage across a changing set of nodes, with no central coordinator. Wrote the node join and leave protocols and successor replication so data survives when nodes drop out, and tested it under simulated node churn.

PythonDistributed Hash TableConsistent Hashing

Attendance Management System

Full-Stack Developerpersonal

Built a full-stack attendance system for a client firm: a Flutter mobile app, a Node.js/Express REST API, and a Next.js admin dashboard, from database schema to production deployment on cPanel shared hosting. Added GPS geofencing with the Haversine formula to confirm a user is on-site, and tied each account to one registered device to stop proxy check-ins. Also built role-aware JWT auth, on-time/warning/late status based on configurable time windows, and automated daily and monthly PDF reports for management.

FlutterNode.jsTypeScriptNext.jsMySQLFirebaseExpresscPanelJWTHaversine formula

Workflow Automation Pipelines

Automation Engineerpersonal

Built a daily tech-news assistant as a scheduled n8n workflow. It scrapes tech RSS feeds, runs the articles through a Gemini node to filter, dedupe, and summarise each story, then emails a formatted digest through Gmail each morning. It skips weekends and sends an alert if a feed goes down. Built a RAG chatbot for employee questions. An n8n workflow indexes new and updated Google Drive documents into a Pinecone vector database using Gemini embeddings; on a query, it retrieves the top-k relevant chunks, passes them to a Gemini node as context, and answers from them, so replies track the current documents without retraining.

n8nGemini AIGoogle DrivePineconeWebhooksGmailRSSVector DatabaseEmbeddingsRAG

Movie Recommendation Lakehouse

Data Engineerpersonal

Built a Bronze/Silver/Gold pipeline on Delta Lake over MovieLens, swapping full-table overwrites for MERGE upserts keyed on user plus movie so reruns stay safe, and scheduled it with Databricks Workflows. Trained an ALS recommender at RMSE 0.88 on a held-out split, tracked in MLflow; Unity Catalog blocked Spark's built-in recommendForAllUsers, so top-N picks come from a window function over predicted ratings, feeding a dashboard with a user filter and a 50-rating minimum on the top-movies chart.

DatabricksPySparkDelta LakeMLflowSQLUnity CatalogALS Recommender

Education

BS Computer Science & Minor Economics

Lahore University of Management SciencesComputer ScienceAug 2023 – May 2027

Coursework

Data StructuresAlgorithmsFoundations of AI/MLDatabasesSoftware EngineeringOOP

A-Level

Lahore Grammar School – JTAug 2021 – May 2023

Coursework

3A, 1BO-Level: 5A*, 2A

Recognition

Awards

Merit Scholarship 100%

LGS JT

Awarded for A-Levels

High Achievers Award: A-Levels

Lahore Grammar School

High Achievers Award: O-Levels

Lahore Grammar School

Hire AbdullahReach out about a role, a contract or a conversation.For recruiters

Abdullah's twin is AI, it can make mistakes.

More people on Kanvis

Browse more people Create your own Kanvis