Abdullah Temel

Experience

• 

  BiletBank

  Apr 2019 - Jul 2020

  Software Engineer

  • Developed Python and .Net applications for Bank reconciliation and fraud detection systems.• Utilized Pandas, Keras, and Matplotlib to enhance system efficiency and accuracy.

• 

  Cimri

  Jul 2020 - Mar 2022

  Software Engineer

  In my role at Cimri, I spearheaded the development of image micro-services for a market price comparison website, managing over 10M products data daily. Additionally, I implemented a push notification system using Akka and Kafka to track market price brochures for millions of users. • Utilised Scala(Akka Http, Akka Streams), Python(Flask), Java(SpringBoot) for development • Managed data with PostgreSQL, Cassandra, and Solr, ElasticSearch• For Events, Utilised Kinesis, SQS, Kafka, Debezium Show less

• 

  Kiwilaunch

  Feb 2022 - Aug 2023

  Founding Software Engineer

  • As a single backend developer Led the design, implementation, and deployment of appointment services, payment processing, and CRM functionalities for Kiwilaunch.• Developed software using .Net 6, Python, MSSQL, ElasticSearch, and AWS services like EKS, ECR, SQS, and SES.• Successfully released the Kiwilaunch product, contributing to the global online booking and scheduling software for service businesses.

• 

  Adevinta

  Aug 2023 - now

  Senior Data Engineer

  Adevinta is a leading online classifieds group, operating 25+ digital marketplaces across Europe and beyond including Marktplaats in the Netherlands.•Led the development and design parts of a GEN-AI service includes serving, tracking, testing and monitoring for ad descriptions as utilising Anthropic models which is rolled out for Android user on Leboncoin in scope of A/B.• Migrated 2 classifier-based products and contributed to migrating 2 more products for automating the ad insertion process serving more than 4 million requests daily.• Designed test and common library templates for classifier-based products to train, serve the models, reducing development time by 70%.• Integrated new version of Marktplaats category recognition model which increase the acceptance rate by 12%.• Designed package size recommender service for reducing shipping parcel label to decrease cost of shipping process. • Implemented Airflow Sensors to check data and trigger notification pipeline. Show less