I build applied AI products with Python, FastAPI, NLP, machine learning, and full-stack deployment.
I'm a Software Engineering student from Karachi focused on AI/ML roles, applied machine learning, and full-stack AI products. My work spans Python, FastAPI, scikit-learn, transformers, NLP, MedGemma, and deployable React/Next.js interfaces.
I build systems where models are useful inside real workflows: phishing detection, clinical triage, manipulation-risk scoring, infrastructure failure prediction, crop-yield forecasting, and skill-gap analysis. The focus is not only training models, but making predictions explainable, usable, and reviewable.
Explainable NLP platform detecting fear, urgency, authority abuse, and polarization with transformer analysis, confidence scoring, and evidence trails.
Multi-layer phishing and malicious URL detection with a Chrome extension, ML models, and live monitoring dashboard.
Hybrid ML engine predicting utility failures across power, water, traffic, and transformer grids with cascading-failure simulation.
Clinical decision intelligence engine ingesting unstructured medical reports, extracting biomarkers, predicting disease risk via Random Forest, and explaining results with SHAP.
Offline-capable clinical triage assistant using MedGemma to evaluate symptoms, vitals, and history with Docker deployment options.
AI-assisted onboarding, payout flows, and risk evaluation with Hedera audit logging and multi-tenant RBAC.
AI-powered platform mapping user skills against job market demands via NLP, generating personalized learning roadmaps with radar charts and heatmaps.
ML models predicting crop yields using soil, climate, and historical data with feature engineering and explainability.
Full MERN healthcare platform: admin, doctor, patient roles, telehealth, billing, Redis caching, and analytics.
ZK-proof payment flows on Starknet L2 with Pedersen commitments and nullifier-tree double-spend prevention.
Interactive narrative simulation with branching flows, live audience voting, configurable voting windows, and dynamic state updates.
Trustless Bitcoin to Hedera swaps using SHA-256 HTLCs with non-custodial claim, refund, and monitoring flows.
3D interactive visualization of GitHub contribution history using React Three Fiber, Zustand state management, and the GitHub API.
Fractionalizes institutional green bonds into tokenized retail units on Hedera with verifiable impact traceability.
Conversion-focused e-commerce storefront with dynamic product management, admin panel, and switchable JSON/MongoDB backend.
University of Karachi
2023 - Present
Aptech Learning
2022 - 2025
Aptech Learning
2023Aptech Learning
2023I'm looking for software engineering roles across full-stack, frontend, backend, and AI-enabled product development. If my work fits your team, let's talk.