Fatima Khalid

Software Engineer & ML Enthusiast Specializing in Predictive Modeling and Full-Stack Automation

Based in: Lahore, PakistanMost recently: Full-Stack Python Developer, Smart A & I Solutions

A Software Engineering student at COMSATS with a strong background in building high-accuracy ML pipelines and production-grade ETL automation. Experienced in full-stack development using Next.js and Python, with a proven track record of delivering data-driven SaaS features and optimizing database systems.

Experience

Full-Stack Python Developer

Smart A & I SolutionsworkJul 2026 – Aug 2026

Summary

Delivery Compliance System (Next.js · TypeScript · Django REST)

What I did

Built a CSV Imports module end-to-end (upload forms, import log tables, batch processing) against 5 backend endpoints with auto-refresh via a lightweight event bus.
Debugged API contract mismatches across Vehicles, Deliveries, and CSV Imports by inspecting live network responses; fixed silent-failure bugs caused by the frontend reading the wrong response shape.
Resolved a 5-part auth issue chain: incorrect URL prefixes, login field mismatch, missing JWT role claims, CORS errors, and a client/server hydration mismatch in a role-based route guard.
Built a reusable Fleet Map with Leaflet and OpenStreetMap - status-colored markers, detail popups, near-real-time updates, and a centralised theme config.

Results

Designed system for a fleet of around 200 vehicles and a few thousand deliveries per day.

Next.jsTypeScriptDjango RESTPythonLeafletOpenStreetMapJWTEvent BusReact

MIS Intern

Shaukat Khanum Memorial Cancer Hospital & Research CentreLahoreinternshipAug 2025 – Aug 2025

Summary

Database (Oracle)

What I did

Queried and maintained Oracle database records across 3 clinical modules, ensuring data consistency for 500+ patient entries.
Optimised SQL reporting queries, reducing manual lookup time by an estimated 40% for the department's weekly data pulls.
Documented table relationships and field definitions for 2 core modules, contributing to the internal data dictionary.

Results

Ensured data consistency for 500+ patient entries.
Reduced manual lookup time by an estimated 40%.

OracleSQL

Skills

Technical

Python

scikit-learn

pandas

Next.js

Jupyter

Git/GitHub

VS Code

NumPy

Seaborn

Selenium

Google Sheets API

Node.js

FastAPI

Django REST

MongoDB

Google Colab

PyCharm

SQL

JavaScript

Matplotlib

MySQL

TensorFlow

Java

n8n

GridSearch

RandomSearch

Fine-tuning Feasibility Analysis

Multimodal Inference

ML Licensing Due Diligence

AbortController

React

Model Evaluation

Multi-threading

Synchronization

Areas of expertise

Leadership

Sponsorship Outreach

Budgeting

Projects

Data Extraction & Form Automation Bot

personal

Production-grade ETL: extracts from Google Sheets (OAuth2), transforms field mappings, auto-populates web forms via Selenium across dynamic page states with modular extract/transform/load functions.

PythonSeleniumGoogle Sheets APIOAuth2

View the code

Tesla Stock Price Analysis & Predictive Modelling

academic

End-to-end ML pipeline on 2,401 trading days (2016-2025): EDA, 10+ technical indicators (RSI, MACD, ATR, ROC, VWAP), IQR outlier removal, normalisation; benchmarked Random Forest, ANN, KNN, Logistic Regression. Random Forest: 94.6% accuracy, 93.5% CV score | ATR 62.3% feature importance, MACD 21.3% | LR: 88.5% accuracy, 1.3% train-test gap.

PythonPandasscikit-learnMatplotlibRandom ForestANNKNNLogistic RegressionGridSearchRandomSearch

View the code

Fitness Club Full-Stack Application

personal

Migrated Node.js/Express to Next.js App Router with clean MongoDB schema abstraction consumed via API routes; CI/CD pipeline deployed to Vercel.

Next.jsNode.jsMongoDBExpressVercelCI/CD

Visit the site

Object-Oriented Chat Application

Developerpersonal

Developed a multi-client chat application using Java to handle concurrent user connections and message persistence.

JavaServerSocket

Final Year Project

personal

Developed a virtual try-on feature with a custom hook to coordinate asynchronous operations for image generation and live color-swatch previews.

ReactAbortControlleruseRef

Ad Spend Optimization

Founder & Developerpersonal

Built a SaaS platform analysing Meta and Google Ads CSV exports across 3 dimensions - placement, age group, and time of day - to identify inefficient spend and recommend budget adjustments. Developed a FastAPI + Next.js application with account creation, CSV validation (up to 20 MB / 500,000 rows), automated analysis, client dashboards, and downloadable PDF reports. Added tenant-isolated data access, configurable report approval, and admin tools for account management; validated the frontend with 169 passing tests plus lint, TypeScript, and production-build checks.

FastAPINext.jsTypeScriptCSV ValidationPDF ReportsLintProduction-build checks

VTON Architecture Evaluation & Feasibility Study

personal

Evaluated multiple pre-trained VTON architectures including IDM-VTON, OOTDiffusion, CatVTON, and AlphaBake for specific downstream tasks. Diagnosed performance gaps related to training data bias in Western garments versus South Asian clothing.

IDM-VTONOOTDiffusionCatVTONAlphaBakeSD 1.5SDXL

Education

BSc Software Engineering

COMSATS University Islamabad, Lahore CampusSoftware EngineeringSep 2023 – Jun 2027

Coursework

Artificial IntelligenceData Structures & AlgorithmsDatabase SystemsOOPPattern RecognitionData Science Fundamentals

Governance & advisory

Team Lead / FYP Group Leader

Academic & FYP Teamother2023 – Present

Led 6 group projects; serving as FYP Group Leader - task delegation, deadlines, technical direction.

Social impact & activities

Sponsorship & Events, Treasurer

CFDSorganizer2023 – Present

Managed 2 events (500-600 attendees); led sponsorship outreach, negotiated external partnerships, and oversaw budgeting.

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

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

More people on Kanvis

Browse more people Create your own Kanvis