Pelin Balci, Msc Industrial Engineer, Machine Learning Engineer

About

I hold a master’s degree in industrial engineering and have a professional background in this field. My postgraduate education was completed at Bilkent University, one of the tops institution in industrial engineering in Turkey. In addition to my academic qualifications, I have supplemented my expertise by taking several online courses in machine learning and deep learning.

At present, I am working as a Sr Lead Machine Learning Engineer at Arçelik Global. My research interests are focused on Retrieval Augmented Generation Systems, Knowledge Graphs, Tiny ML, Natural Language Processing, Large Language Models and Reinforcement Learning. These are particularly dynamic areas of study, and I am passionate about advancing my understanding of them.

I am using Python as my primary programming language for my various projects. Additionally, I strive to enhance my coding skills by utilizing LeetCode. My objective is to make meaningful contributions to the data science community in Turkey. Recently, I joined AYA (Açık Yazılım Ağı), an organization committed to helping those impacted by the devastating earthquake in Kahramanmaraş. I have worked as a lead on data labeling for NER and Intent Classification Models. I’m also giving mentorship on the Geleceğini Kuran Genç Kadınlar (Young Women Establishing Their Future) project to support young womens who are neither in education nor in employment.

As a self-learner, I firmly believe in the importance of open-source software. I embrace this philosophy by sharing my codes on GitHub and publishing posts on my website and Medium. If you have any questions about my projects or wish to engage in an open-source initiative, please do not hesitate to contact me :)

Contact

Key Skills

Operations Research, Mathematical Modeling, Simulation, Data Science, Machine Learning, Deep Learning, Tiny Machine Learning, Reinforcement Learning, NLP, LLM, Prompt Engineering, Transformers, NER Models

Software Knowledge

Python for Data Science (Pandas, Numpy, Simply Scikit-learn, Matplotlib, PyTorch, Simpy, TensorFlow, TensorFlow Lite, Streamlit, llama-index, langchain), SQL, GAMS, Power BI

Personal Projects

Live Apps: Apps Page

Work Experiences

  • 2024 - Arçelik Global, Senior Lead Machine Learning Engineer
    • Text Classification & Training Pipeline & AWS deployment
    • Retrieval Augmented Generation Project: Knowledge Graphs, Metadata Extraction
  • 2023 SabancıDx, Lead Data Scientist/Machine Learning Engineer
    • Compilation of Güler Sabancı’s speech at the Sabancı University Graduation Ceremony using Azure OpenAI service and GPT models Sabancı University News YouTube
    • Web/App Automation Product with NLP and GPT Models
    • Air Handling Unit Energy Optimization Product with Reinforcement Learning
  • 2022 SabancıDx, Sr. Data Scientist/Machine Learning Engineer
    • Demand prediction for retail company
      • The aim of this project is to predict the demand from sub-parties and turn the prediction into order amount according to the lead time and safety stocks
      • My responsibilities are re-framing the project scope, detailing the tasks, making code reviews, turning the initial codes into production
  • 2020 - 2022 SabancıDX, Data Scientist/Machine Learning Engineer
    • Prediction of breakage in the production
      • The aim of this project is to recommend sensor values which minimizes the breakage in the final product for industry. XGBoost Regression is used to make break prediction, a recommendation system is implemented which suggests the best sensor values in the factory.
      • Model: XGBoost and Linear Regression
    • IOT Project: Air Quality IOT Product
      • Building an IOT Product which measures the quality of data for offices.
      • Controlling the measurements with Power BI & Python
    • Prediction of System Direction for an Energy Company
      • The aim of this project is predicting the hourly energy imbalance direction for the upcoming day.
      • Model: Neural Network Model is implemented with PyTorch
      • Deployment: Databricks. Our model is being retrained every week and send a prediction mail to custome
      • Tools: Python, Databricks, SQL, Power BI
    • Delivery Optimization for a Fast-Food Company
      • The aim of this project is improving delivery process and optimize shifts
      • Simulation for delivery process: I’ve written Simulation environment with Simpy library which shows the order’s assignments and delivery times. Simulation is used to show the current state and the improvement of our new approach before live test.
      • Shift optimization: Targeted hourly driver numbers are calculated based on order distributions. The working hours of full and part-time employees are determined by a mathematical model based on the targeted employees. The objective function is minimizing the difference between the target driver and the assigned driver. Simulation environment is used to analyze the proposed number of employees.
      • Tools: Python, SQL, Gurobi, Simpy, Power BI
  • 2019 – 2020 Adphorus a Sojern Company, Sr. Data Analyst
    • Reporting the performance of algorithms
      • Send status reports by email
      • Tools: Facebook API, Python, SQL
    • A/B Testing
      • Sampling Algorithms, Statistical Tests
      • Facebook Digital Marketing
  • 2017 – 2019 Enerjisa Sales, Commodity Portfolio Management Specialist
    • Mark to Market Reporting System Project
      • Design the reporting tool with IT Department
      • Keep versions to make comparisons with budgets
      • Make User Tests
      • Control Profit and Loss Calculations
      • Prepare Daily, Weekly, Yearly Reports
    • Process development
    • Customer based demand analysis
    • Prepare hedging strategies in terms of electricity in MW and dollar
    • Calculate energy prices for customers considering dollar, market conditions and risk premiums
    • Manage short positions
  • 2015 - 2017 Enerjisa Trading, Portfolio Optimization Specialist
    • Power Generation Optimization Tool Project:
      • Design the optimization tool with IT Department
      • Keep versions to calculate the difference between planned and actual generation in terms of production and prices
      • Make User Tests
      • Control Profit and Loss Calculations
      • Prepare Daily, Weekly, Yearly Reports for locked and forward P&L
    • Optimizing, planning and simulating long term generation of power
    • Make budget of all power plants for upcoming years
    • Prepare hedging strategies in terms of electricity in MW
    • Manage long positions
    • Create P&L from delta hedging (buy and sell electricity based on market conditions)
  • 2013 - 2015 MilSOFT Software Technologies Inc., Project Management Specialist
    • Software project management (CMMI5, Agile and Scrum Methods)
    • Risk management for known unknown bugs and milestones for projects
    • Statistical analysis for HR Planning
    • Earned value analysis
  • 2010 - 2012 Bilkent University Industrial Engineering Department, Teaching Assistant

Education

  • 2010 - 2013 Bilkent University Graduate School of Engineering and Science
  • 2005 - 2010 Gazi University Engineering and Architecture Faculty
    • Industrial Engineering, Bachelor of Science GPA: 3.44 / 4.00

Presentation

P. Balcı, B. Tansel “Analysis of Locations of Existing Fire Stations in Ankara in Comparison to Optimized Locations.,” International IIE Conference, The Global Reach of Industrial Engineering, Istanbul / Turkey, June 26 - 28, 2013

Awards

  • Third Prize: Advanced Data Analytics, Data Scientist Track, Sabancı University, 2018
  • Third Prize: Brabant Water Demand Prediction Problem, JADS 7th Data Challenge Week, 2019

Scholarships

  • Bilkent University Graduate School of Engineering and Science Scholarship, 2010 – 2012
  • TÜBİTAK Graduate Scholarship, 2010 – 2012

Trainings

Deep Learning

  • Building Applications with Vector Databases, deeplearning.ai, 2024, Freely Available
  • Google Cloud Lectures, 2023 certificates
  • Finetuning Large Language Models, deeplearning.ai, 2023, Freely Available
  • Introduction to Large Language Models, Google Cloud, Coursera, 2023, certificate
  • Generative AI with Large Language Models, Deep Learning AI, Coursera, 2023, Free Version
  • ChatGPT Teach-Out, Michigan University, Coursera, 2023 Free Version
  • Natural Language Processing with Classification and Vector Spaces, Deep Learning AI, Coursera, 2022 Free Version
  • Sequence Models, Deep Learning AI, Coursera,2022, certificate
  • Custom Models, Layers, and Loss Functions with TensorFlow, Deep Learning AI, Coursera, 2022 certificate
  • Convolutional Neural Networks in TensorFlow, Deep Learning AI, Coursera, 2021 certificate
  • Introduction to TensorFlow for Artificial Intelligence, Machine Learning, and Deep Learning, Deep Learning AI, Coursera, 2021 certificate
  • CS231n Convolutional Neural Networks for Visual Recognition, Standford University, 2020github_notes
  • Intro to Deep Learning with PyTorch, Udacity, 2020 github_notes

Embedded Machine Learning

  • Applications of TinyML, HarvardX, Edx, 2022 Free Version
  • Fundamentals of TinyML, HarvardX, Edx, 2022 Free Version
  • Introduction to Embedded Machine Learning, Edge Impulse, Coursera, 2021 certificate

Machine Learning

  • Machine Learning Specialization, Deep Learning AI, Coursera,2022 certificate
  • Unsupervised Learning, Recommenders, Reinforcement Learning, Deep Learning AI, Coursera,2022 certificate
  • AWS Machine Learning Foundation Course, Udacity, 2020 github_notes
  • Machine Learning Crash Course with TensorFlow API, Google, 2020 github_notes
  • Intro to Machine Learning, Udacity, 2020 github_notes
  • Bayesian Machine Learning via Python: A/B Testing, Udemy, 2020 github_notes, certificate
  • Advanced Data Analytics, Data Scientist Track, Sabancı University, 2018

Other

  • The Science of Stem Cells, American Museum of National History, Coursera, 2023, certificate
  • Self Awareness and the Effective Leader, Rice University, Coursera, 2022 certificate
  • Discrete Optimization, The University of Melbourne, Coursera, 2020 certificate
  • Bertelsmann Data Analyst Nano Degree with Scholarship, 2018 – 2019 certificate
  • Google Android Basics Nano Degree with Scholarship, 2017 – 2018 certificate
  • Intermediate Python for Data Science Course, DataCamp, 2017 certificate
  • Intro to Python for Data Science Course, DataCamp, 2017 certificate
  • Data Science Orientation by Microsoft, 2017 certificate

Foreign Languages

English (Advanced), Spanish (Beginner), German (Beginner)

Interests and Personality Characteristics

  • Minimalism, Cycling, Lindy Hop Dance, Solo Jazz Dance, Chess
  • TÜBİTAK Formula - G Solar Car Race, G-Mobil 2 Sponsor Team Member, 2008 – 2009
  • Gazi University Technology Club, Charter Member and Member of the Board, 2007 – 2009
  • Self-learner, self-awareness, a good team member, creative and analytical thinking, good communication skills

Badges & Projects

Edge Impulse badge
Movie Recommendation System-1
Movie Recommendation System-2