Sydney, NSW

Shaheer Alvi

Data Science & AI Engineer

About

I’m a penultimate-year Data Science Engineering student at UTS, carrying a Distinction WAM.

I build machine learning systems and data pipelines end to end — from raw, noisy input through to something that actually runs.

I’m drawn to applied ML inside real financial systems, particularly fraud detection.

Optik

Client: Royal North Shore Hospital

Built and validated an end-to-end pipeline extracting respiratory impedance from a noisy real-time stream.
Covered the pipeline with ~65 automated tests.
Worked as part of a multidisciplinary team.

Team of 6 · UTS · Backend & systems integration

AI-Driven Digital Payment Platform

Built the transaction lifecycle — hold, approve, flag, block — on top of a non-custodial buffer.
Integrated a fraud-scoring model and an 8-check KYC flow into the transaction path.
Drove a SQLite migration to hit a sub-500ms target, and validated 16 API endpoints.

Python · Flask · SQLAlchemy · SQLite · REST APIs

UTS 42028

Real-Time Road Sign Detection

Trained YOLOv8s to 0.714 mAP@0.5 on test and 0.858 on validation, across the 50-class, 24k-image TT100K benchmark.
Built a custom 5-block CNN (4.86M params) for patch classification at 99.69% test accuracy.
Trained the CNN on AWS SageMaker (T4).

Python · PyTorch · YOLOv8 · OpenCV · Streamlit · AWS SageMaker

Solo · UTS 41001

Cloud-Deployed Web App

Deployed a Flask app solo on AWS Elastic Beanstalk.
Designed a custom VPC with public and private subnets, Multi-AZ RDS MySQL, Auto Scaling and SNS.
Debugged a driver-level authentication failure through to root cause.

Python · Flask · AWS (EB, VPC, RDS, Auto Scaling, SNS) · MySQL

Data analysis

City of Sydney

Python · SQL · Power BI · Excel

Training, evaluating and tuning models for vision and risk problems.

Turning raw, multi-source data into something a model or a person can use.

Building the services and infrastructure that put a model into production.

Sydney, NSW