Muhammad Roshaan · Backend & Data Engineer

Backend data,
engineered to ship.

Final-year CS at SSUET building systems that survive contention: distributed locks, ETL pipelines at 540K+ rows, and real-time ML inference you can audit.

GitHub Résumé Email LinkedIn

540K+

records through the ETL pipeline

0.9995

ROC AUC on a 1.27M-sample fraud set

6

systems shipped end-to-end

3.81

CGPA · BS Computer Science

Python•FastAPI•PostgreSQL•Redis•Celery•Docker•SQLAlchemy•MongoDB•Supabase•scikit-learn•XGBoost•LightGBM•Next.js•Flutter•Java / Spring Boot•Power BI•Python•FastAPI•PostgreSQL•Redis•Celery•Docker•SQLAlchemy•MongoDB•Supabase•scikit-learn•XGBoost•LightGBM•Next.js•Flutter•Java / Spring Boot•Power BI•

01 — Selected work

Systems, chosen to explain themselves

Case studies over links. Each project has a page walking through the problem, the approach, and what actually changed.

01 / Backend · Distributed Systems

SeatVault

Backend system for high-contention reservations with atomic inventory control, Redis distributed locking, PostgreSQL row locks, and idempotent payment-style workflows.

FastAPI
PostgreSQL
Redis
Celery
Docker
02 / Data Engineering · ETL

Enterprise ETL & Inventory Pipeline

A data pipeline that ingests inventory, transforms and validates records, bulk-loads structured output into Supabase/PostgreSQL, and emits alerts for operational conditions.

FastAPI
PostgreSQL
Supabase
SQLAlchemy
Celery
Redis
03 / AI/ML · MLOps

Sentinel

Fraud-detection service combining real-time event processing, ML inference via a max-vote ensemble, explainability, and operational monitoring.

FastAPI
scikit-learn
XGBoost
LightGBM
Redis
PostgreSQL
Hopsworks
04 / Full Stack · Offline-First

BooknScore

Offline-first tape-ball cricket scoring with player and team analytics, authentication, synchronization, and a companion Python + Gemini AI service.

Flutter
Supabase
PostgreSQL
SQLite
Python
Gemini
05 / Backend · Geospatial

Geospatial Fleet Dispatch API

Serverless real-time fleet dispatch ingesting driver coordinates and mapping nearby vehicles via geospatial database queries, surfaced in a Next.js command center over WebSockets.

FastAPI
MongoDB
Cloudflare Workers
Next.js
WebSockets
06 / Backend · Multi-Tenant

Dealer & Vehicle Inventory Module

Java Spring Boot modular monolith with tenant-aware data isolation, custom query filtering, and role-based access control.

Java
Spring Boot
Spring Data JPA
Maven

02 — Capabilities

The stack I reach for

Languages, frameworks, and runtimes, grouped by where they earn their keep in my work.

01

Languages

Where the systems are written.

Python
Java
SQL
02

Backend & APIs

Services, contracts, and background work.

FastAPI
SQLAlchemy
Celery
Pydantic
REST
WebSockets
03

Databases

Relational, document, and in-memory stores.

PostgreSQL
MySQL
SQLite
MongoDB
Redis
Supabase
04

Data & ML

Pipelines, models, and explainable inference.

Pandas
NumPy
scikit-learn
XGBoost
LightGBM
Hopsworks
Power BI
05

Tools & Platforms

The workspace.

Git
Docker
Jupyter

03 — Path

Where it's being applied

A structured setting to turn what YouTube-style learning into something repeatable.

APR—JUN 2026

Data Science Intern @ 10Pearls Pakistan

10P Shine Internship Program · NASTP · Remote

  • Built and evaluated machine learning workflows for practical data science tasks.
  • Applied data preparation, feature engineering, model evaluation, and experiment tracking concepts.
  • Worked with Python data and machine learning tooling in a structured internship environment.
Python
Pandas
NumPy
scikit-learn
XGBoost
LightGBM
Hopsworks
Jupyter

04 — Contact

Let's build something that survives.

Open to backend and data engineering internships and junior roles. If that's you, my inbox is open.

Muhammadroshaan1127@gmail.com
GitHubLinkedInEmailRésumé

© 2026 Muhammad Roshaan · Built with Next.js, Tailwind & Framer Motion