Burak Bastem

Burak Bastem

Co-Founder & CTO
at Imperium

burakbastem@hotmail.com

Last updated: February 4, 2024

Bio

Burak Bastem is a software engineer specialized in software development (SD), machine learning (ML) and high performance computing (HPC). He studied, applied and developed state-of-the-art ML solutions and pipelines for fraud detection, a challenging problem due to its highly unbalanced nature. His experience in HPC involved developing GPU and CPU runtime libraries, optimizing multi-process communication and accelerating applications. He now focuses on software development for construction project management.

Experience

Co-Founder & CTO
November 2023 - present
Imperium
Turkiye C#,JavaScript, .NET, React, React Native
  • Providing software for construction project management.

Senior Software Engineer
April 2022 - October 2023
ASML via Luxoft
Eindhoven, NL Python, Data Definition Format (DDF), MATLAB, Linux
  • Developed and integrated CPD (calibration-performance-diagnostic) tests for EXE5000 POB (projection optic box).

Senior Software Engineer
February 2022 - April 2022
Blue Cloud
US - remote ML, Python
  • Started a POC for JLL to predict whether work orders would be completed on time with machine learning.

Senior R&D Engineer
August 2020 - February 2022
Yapı Kredi Teknoloji - Applied Data Science Team
İstanbul, Turkiye ML, Fraud Detection, Big Data, Parallel Programming, Computer Vision, Python, Java, SQL, Docker, MongoDB, Redis, Hive, Spark, Pandas, scikit-learn, Linux
  • Worked on fraud focused machine learning projects: cross-channel fraud detection and employee fraud detection. Due to difficulty of the problem, projects were mainly research-focused and enriched with academic collaborations from respectable universities. The projects also involved HPC and big data solutions because of their data size.
  • While studied, applied and developed state-of-the-art machine learning solutions in the research phase, implemented and integrated pipelines (backend solutions) in the production phase.
  • Developed a proof-of-concept computer vision project about tracking social distance in bank branches from camera footage.

Expert R&D Engineer
July 2018 - August 2020
Yapı Kredi Teknoloji - Applied Data Science Team
İstanbul, Turkiye ML, Fraud Detection, Python, Java, SQL, Redis, Pandas, scikit-learn, TensorFlow, Linux
  • Developed machine learning models for credit card fraud detection which involves providing solutions for extremely unbalanced class ratio and concept drift.
  • Achieved better performance than proprietary model.
  • Developed and integrated a pipeline to train models automatically and score credit card transactions in real-time (<10 ms).
  • Represented the company and fulfilled the responsibilities as a data provider in European Data Incubator (EDI), a European Union H2020 project.

Research Assistant
June 2015 - April 2019
Koç University Parcorelab
İstanbul, Turkiye HPC, Parallel Programming, GPU Programming, C++, CUDA, OpenACC, MPI, OpenMP, Pthreads, Linux
  • Implemented TiDA-C++, tiling based multi-threaded programming model to increase cache performance with data locality and manage parallelism. TiDA achieves up to 2.10x speedup over OpenMP and resulted in publication at ISC’16.
  • Designed and developed TiDA-acc. TiDA-acc is a tiling-based GPU programming model which manages distinct address spaces, successfully hides transfer latency between CPU and GPU, automatically generates GPU code, handles cases where there is no sufficient GPU memory and is published at ICPP’17.
  • Integrated TiDA-acc to SMC, a combustion simulation consisting of approximately 10,000 lines of code, and presented an article and a poster about the study at BAŞARIM’17, Turkiye’s national HPC conference.
  • Developed a tiling-based high-level asynchronous programming model for GPU clusters using TiDA-acc. On top of what TiDA-acc offers, it utilizes all GPUs (and CPUs) in the system, overlaps any type of communication with computation and achieves good speedup. The study is published at HPCAsia’20.

Teaching Assistant
Sept 2016 - July 2018
Koç University
İstanbul, Turkiye
  • Assisted in teaching, supervision and assessment of Computer Architecture, Operating Systems and Parallel Programming courses.
  • Received Teaching Assistant Training Certificate.

Research Assistant
July - Sept 2016
Lawrence Berkeley National Laboratory
Berkey, CA, US HPC, Parallel Programming, GPU Programming, C++, CUDA, OpenACC, Linux
  • Studied existing GPU execution models and SMC implemented with AMReX, an AMR framework from Berkeley Lab.
  • Implemented GPU version of SMC kernels.
  • Collaborated on TiDA-acc design.

Software Engineering Intern
August 2015
Vestel R&D
İstanbul, Turkiye Java, Android
  • Developed proof-of-concept applications for Android platform.

Software Engineering Intern
June - August 2014
Netaş
İstanbul, Turkiye Java
  • Analyzed and configured static code analysis tools to be integrated to continuous delivery project.

Education

Master of Science, Computer Science and Engineering
Sept 2016 - Apr 2019
Koç University
İstanbul, Turkiye

Thesis: Tiling-Based Programming Model for GPU Clusters Targeting Structured Grids
GPA: 3.88 / 4.00


Bachelor of Science, Computer Engineering
Feb 2012 - Feb 2016
Koç University
İstanbul, Turkiye

GPA: 3.18 / 4.00


Bachelor of Science, Industrial Engineering - Double Major
Feb 2015 - June 2016
Koç University
İstanbul, Turkiye

GPA: 3.18 / 4.00


Track Certificate Program, Software Engineering
Feb 2014 - Feb 2016
Koç University
İstanbul, Turkiye

GPA: 3.54 / 4.00


English Language Program
Sept 2010 - Jan 2012
Koç University
İstanbul, Turkiye

Publications

Conference Proceedings

Burak Bastem, Didem Unat, “Tiling-Based Programming Model for Structured Grids on GPU Clusters”, International Conference on High Performance Computing in Asia-Pacific Region (HPCAsia), Fukuoka, Japan, January 2020

Burak Bastem, Didem Unat, Weiqun Zhang, Ann Almgren, John Shalf, “Overlapping Data Transfers with Computation on GPU with Tiles”, International Conference on Parallel Processing (ICPP), Bristol, United Kingdom, August 2017

Didem Unat, Tan Nguyen, Weiqun Zhang, Muhammed Nufail Farooqi, Burak Bastem, George Michelogiannakis, Ann Almgren, John Shalf, “TiDA: High-Level Programming Abstractions for Data Locality Management”, International Supercomputing Conference (ISC), Frankfurt, Germany, June 2016

Honors and Awards

Research Project Scholarship
Sept 2016 - July 2018
Prof. Didem Unat via TÜBİTAK

Half Merit Scholarship
2010 - 2016
Koç University, Istanbul, Turkiye

Second Best Senior Project
2016
Koç University, Istanbul, Turkiye

Dean's Honor Roll
Fall 2014, Fall 2015, Spring 2016
Koç University, Istanbul, Turkiye

Vehbi Koç Scholar
Spring 2014, Spring 2015
Koç University, Istanbul, Turkiye

Activities

Employee Representative, Occupational Health and Safety Board, Yapı Kredi Teknoloji
2020 - 2022

Speaker, TRAI Meet-Up #45
May 2021

Speaker, EuroCC@Turkey HPC Winter School
Feb 2021

Speaker, HPCAsia 2020
Jan 2020
  • Presented the HPCAsia paper.
  • Received Certificate of Attendence.

Speaker and Volunteer, BAŞARIM 2017
Sept 2017
  • BAŞARIM is Turkiye's National High Performance Computing Conference.
  • Presented the BAŞARIM paper and poster.
  • Volunteered in organizing and running the conference.
  • Received Certificate of Attendence.

Speaker, ICPP 2017
August 2017

Member, Koç University Computer Club
2010 - 2013
  • Coordinated and participated in workshops and seminars.

Member, Koç University Marketing Club
2010 - 2011
  • Participated in seminars and career talks with top executives of reputable companies.
  • Received BusinessPort Certificate of Attendence.

Skills

Turkish: Native

English: Fluent
  • IELTS Academic: 7

  • KAPLAN International Certificate of Achievement

    Summer 2012

    New York City, NY, US

New


C#
JavaScript
.NET
React
React Native

Primary


Python
C++
Java
SQL

Docker
MongoDB
Redis
Hive
Spark
Pandas
scikit-learn
TensorFlow

CUDA
OpenACC
MPI
OpenMP
Pthreads

Linux


Secondary


HTML
CSS
PHP
Julia
C
Objective-C
MATLAB

OpenCV
OpenGL
WordPress
WooCommerce

iOS
Android