Melina
Behzadi Nejad

Led and contributed to cross-functional teams alongside designers, product managers, and developers to ship production-ready applications on time. Experienced in full Agile/SCRUM workflows — sprint planning, backlog refinement, retrospectives — plus CI/CD pipelines and containerized deployments with Docker and Azure.

Built microservices systems for platforms like the Event Booking Platform and TripSphere — independent services for Events, Bookings, Rooms, Users, and Approvals, all orchestrated via Docker. Migrating from monolithic to microservices cut system response time by 50% and dramatically improved maintainability and data integrity.

Built polished iOS apps including PursePal (expense tracking with real-time Firebase sync) and Lumina (a full-featured WhatsApp-style messenger) using Swift, SwiftUI, MVVM, and CoreData. I integrate encryption, secure auth, and cloud storage as non-negotiables — because user trust is everything in mobile.

Data is only useful when it tells a story — and that's exactly what I build systems for. I design and optimize relational databases using SQL, MySQL, PostgreSQL, and MongoDB, with a focus on data integrity and query performance. My SQL Database for Clinic Management is a fully normalized patient records system with complex queries, referral tracking, and security-first design. I write analytical queries that surface insights — expense patterns in PursePal, booking trends in TripSphere, or employee records in HR platforms. Python + SQL together is where I really come alive.

I built a Python anomaly detection pipeline at Javada Services that flagged unusual patterns in operational data — catching things that rule-based systems would miss entirely. At Sol Café, I developed demand forecasting models that helped the team plan staffing and inventory more accurately. Outside of work, I've trained a house price prediction model on Kaggle using feature engineering and ensemble methods, and built a CIFAR-10 image classifier with a convolutional neural network in TensorFlow. I use scikit-learn, pandas, and Jupyter as my daily toolkit — and I genuinely enjoy the process of turning messy data into something that actually predicts and explains.