Zaeem Asvat Email & Phone Number

Johannesburg Metropolitan Area

Discovery Health (August 2023 - present) - Key full stack developer for the new Discovery Health application. Technologies I work with include Java Spring Boot, Angular and TypeScript. - Implemented various new backend features, enhancements and fixes for the virtual consultation component of the app which allows patients and doctors to have virtual face-to-face consultations in real time - Implemented various frontend features for the application, including the Personal Health Pathways and Family Consent, as well as many fixes on the application Vodacom CRM Team (October 2022 - July 2023)- Re-engineered various core Vodacom services from an older Java legacy system to a modern stack which included AWS Lambda, SQS, NodeJS, TypeScript and React. Involved analyzing the legacy services and implementing new versions into the new system, with improvements to efficiency and maintainability. - Designed a cleaner and more maintainable implementation for our new services, which my team adopted and refactored to. - Designed and implemented a critical architectural change to our solution using AWS Lambda, SQS and ElastiCache to improve the scalability of our new system to handle massive loads of processing, as desired by our users. Data Science team (Jan 2022 - Sep 2022)- Implemented various natural language understanding models to predict BBD applicant success probability based off their resumes - Key member on a team that developed a computer vision solution to detect birds from video footage of a chicken farm in Cape Town in real-time. The locations of birds detected get sent to a laser which deters birds away from the farm, to avoid spreading disease to the chickens

South Africa

Android Application to assist retail stores with automatic stock-taking and inventory management ● Worked on training models for localizing and classifying products in real world store shelves captured by a mobile phone camera ● Carried out extensive complexity testing for AI models with very large numbers of classes, and designed an efficient and scalable solution for mobile devices in terms of storage size, RAM consumption and inference time ● Built a Kotlin Android application backend with which I integrated product localization and classification neural networks using TensorFlow Lite, PyTorch Mobile, and a Room database ● Built, trained, optimized and integrated convolutional neural network for correcting camera perspective distortions from the images taken of real-world shelves from a mobile device. Also developed an API to conveniently correct the perspective of real-world store shelves as a means of acquiring ground truth data for training the model Email Attachment Classification (Allan Gray):● Developed software to classify documents attached to emails as one of the following categories: Bank Statement, Proof of Payment, Death Certificate, ID Book, Driver's License using machine learning algorithms and convolutional neural networksID Verification (FNB) ● Built and trained convolutional neural networks to read passport numbers on ID Books in the form of dotted-text for verification. Applied image processing algorithms to improve visual readability of dotted text to ease classification for the neural network ● Holographic watermark detection on ID Books

South Africa

As part of my duties as a postgraduate student at the university, I tutored 2nd year computer science students for the following courses: - Algorithm Analysis - Computer Networks

Johannesburg, Gauteng, South Africa

Bachelor Of Science Honours, Computer Science

Bachelor Of Science - Bsc, Computer Science

Education record

Primary School and High School

Nursery School, Grade R And Grade One