ML Engineer & Full-Stack Web Developer
Shawal Khan
I build machine learning systems and full-stack web applications, from healthcare-focused ML research to deployed tools like PDF Hamster.
View Work →Selected Work
Machine Learning
Loan Default Prediction
Logistic Regression + Random Forest pipeline with SMOTE class balancing, ROC-AUC ~0.845. Deployed on Streamlit Cloud.
- Python
- scikit-learn
- Streamlit
Web App
Loan Default Risk Console
Productionized version of the Loan Default Prediction model: a FastAPI + JWT REST API with MongoDB-backed prediction history, a React dashboard, and CI/CD, replacing the original Streamlit demo.
- FastAPI
- React
- MongoDB
Machine Learning
Seizure Prediction
Epileptic seizure prediction with a custom Diff-Logic (Differentiable Logic Gate Network) implementation that beat MLP baselines at every parameter budget.
- Python
- PyTorch
- scikit-learn
Web App
PDF Hamster
Flask web app with a set of PDF conversion tools, deployed at pdfhamster.com.
- Flask
- Python
Final Year Project
MASS: Multi-Agent Analysis and Synthesis System
A multi-agent pipeline (Intake, Researcher, Analyst, Auditor, Orchestrator) that turns a raw business case into an audited, evidence-grounded diagnostic report, catching fabricated statistics and unsupported rankings before they reach a reader. Being presented at a hackathon.
- Python
- Multi-Agent
- ChromaDB
Web App
PitchDeck.live
AI pitch deck builder: describe a startup in one line and it generates a full deck outline via the Gemini API.
- React
- TypeScript
- Gemini API
Web App
Exoplanet Explorer
A navigable 3D map of confirmed exoplanets built from real NASA Exoplanet Archive data: 36 planets across 27 systems, positioned by real coordinates with real orbital periods, radii, and masses.
- React Three Fiber
- Three.js
- NASA API
Machine Learning
LLaMA-XR Extended
Reproduction and extension of a 2026 radiology report-generation paper (LLaMA + QLoRA), implemented from scratch since the authors released no code, then extended with the clinical-accuracy eval and hallucination checks their own limitations section calls for.
- Python
- PyTorch
- LLM Fine-tuning
Web App
LaunchMate
AI project idea generator for students: pick a course and difficulty, and Gemini returns a full spec, tools, file structure, build steps, and learning outcomes. Flask backend with Supabase magic-link auth and saved project history.
- Flask
- Gemini API
- Supabase
About
I'm a final-year student working across machine learning and full-stack web development, from healthcare-focused ML research to deployed tools people actually use.
Machine Learning
- Python
- scikit-learn
- PyTorch
- SMOTE
- Streamlit
Web Development
- TypeScript
- Next.js
- React
- Tailwind CSS
- Flask
Tools & Data
- Git
- GitHub
- ChromaDB
Interests
Healthcare AI & NLP
Focused on applying machine learning to healthcare problems, from predictive modeling to natural language processing.
Interests
Full-Stack Web Development
Building deployed web applications end to end, from Flask tools like PDF Hamster to this portfolio itself, built with Next.js and TypeScript.
Experience
Machine Learning Intern
Code Club
May 2026 - June 2026
Member of SCALE
QEC-IMSciences
February 2026 - Present
Co-Lead
MLSA Peshawar
March 2025 - February 2026
President
IM Computing and Innovation Society (IMCIS)
February 2025 - Present
PR and Outreach Team Member
MLSA Peshawar
February 2025 - March 2025
Education
Bachelor's degree, Computer Science
Institute of Management Sciences (IM|Sciences)
September 2023 - July 2027
Pre-Medicine / Pre-Medical Studies
KSKCCS
April 2018 - July 2022
GitHub Activity
539
contributions in the last year
Last commit Refine case study results and scrollbar on codefolk, 10/10/2026
Updates
Final Year Project · 2026
MASS: Multi-Agent Analysis and Synthesis System
A multi-agent pipeline (Intake, Researcher, Analyst, Auditor, Orchestrator) that turns a raw business case into an audited, evidence-grounded diagnostic report, catching fabricated statistics and unsupported rankings before they reach a reader. Being presented at a hackathon.
Read more →Testimonials
Coming soon
Gathering feedback from people I’ve worked with. Check back soon.
Let’s work together
Open to freelance and collaboration opportunities.
Get in touch