Hello!
This is Sara, Machine Learning Developer/ Market Fundamentals and Data Analyst at Alberta Electric Systems Operation (AESO), and a proud alumni of the University of Calgary.
Proficient in designing, training, and deploying machine learning models, including neural networks and deep learning architectures, for diverse applications such as object detection, natural language processing (NLP), and resource/Energy usage forecasting. Adept at creating efficient data processing pipelines for complex datasets, including 3D point cloud data, and extracting meaningful features for machine learning tasks. Hands-on experience in developing end-to-end ML pipelines, from data preprocessing to model deployment, utilizing Azure and Docker for scalability and efficiency.
Formerly...
I graduated from the University of Calgary with a Master's degree of Electrical and Software Engineering in May 2023, meanwhile I took an Internship at AltaML to practice my skills in an industrial level as a Machine Learning Developer and I was able to show my skilss in many aspects. Befor that, I was a full-time AI Specialist at Fanap Co., where I worked on the prediction of resources on a cloud platform.
I'm beyond fortunate to work with many folks at AESO, AltaML and Fanap, where I had the opportunity to learn and grow.
AESO
Nov. 2023 – Present | Machine Learning Developer - Market Fundamentals and Data Analyst

AESO, AESO is the independent operator of Alberta's electric system. We manage and plan the province's power grid and energy markets. Us in Forecast Analytics team at the AESO, provide the Grid with a long term load forecast of every single area in every level to help the reliability of the system. I specifically am involved in Designing, developing and deploying a big data machine learning operation solution for long-term load forecasting (LTO project). Provide technical support and guidance to team members for machine learning projects. Engineered a highly efficient Data Processing pipeline for analyzing and extracting features from million records of historical data, enhancing the performance of machine learning models. Reformat and improved the Transmission Rate Projection (TRP) project calculations. Transferred all the processes to python and modernized it.
  • Built a big data application from the ground up using Python, incorporating robust design principles
  • Improved team collaboration and deployment efficiency by implementing CI/CD workflows on local servers.
  • Utilized Celery and RabbitMQ to enable parallel task management in web applications, improving scalability and performance.
  • Developed innovative methodologies to achieve lower MAPE, advancing forecasting accuracy.
  • Built and deployed a fully functional Flask application using Gunicorn, ensuring reliable server-side operations.
  • Dockerized a Flask application for streamlined deployment on local servers.
  • Developed a PowerBI project to demonstrate all the rates on TRP project.
AltaML
Jan. 2023 – Ma. 2023 | Machine Learning Developer - Intern

ALtaML, Is an AI company that brings Ideation to Evolution, and Accelerates Your AI Journey. As an Associate ML developer I Collaborate with cross-functional teams to develop and implement machine learning models and pipelines. Conduct research and stay up to date with the latest advancements in machine learning and natural language processing. Provide technical support and guidance to team members for machine learning projects. Engineered a highly efficient Data Processing pipeline for analyzing and extracting features from 3D point cloud data, enhancing the performance of machine learning models. Participated in the development of comprehensive unit tests for an ML-Ops Boot Camp, ensuring the reliability and scalability of machine learning workflows through GitHub actions.
  • Played primary development role in data science team in developing and implementing machine learning pipeline object detection on 3D point cloud data, resulting in a 20% increase in accuracy.
  • Spearheaded the development of a Heart Disease Detection Project, where I ranked first among a cohort of 40 in completing a rigorous technical ML Boot Camp. Successfully implemented GitHub actions for an ML Bootcamp project, reducing the reviewing time by more than 50%.
  • Utilized Azure to create automated model training and deployment workflows, achieving a remarkable 50% reduction in time required for model deployment.
  • Developed novel algorithm for sentiment analysis, achieving sentiment prediction accuracy of 90%.
  • Collaborated on analysis and presentation of findings for clients in 3D point cloud data project, contributing to the scientific community's understanding of the subject.
Fanap Co.
Sep. 2019 – Oct. 2020 | Artificial Intelligence Specialist

SAKKU Cloud Platform, is a Docker container-based cloud platform! There, I Researched and implemented machine learning algorithms to accurately forecast resource usage for the SAKKU Platform. Developed Restful-API modules to efficiently collect required data for machine learning development. Designed and implemented file search feature to improve user navigation and productivity. Collaborated with cross-functional teams to ensure seamless integration of ML algorithms and Restful-API modules. Conducted regular testing and debugging to maintain the functionality and efficiency of the ML algorithms and Restful-API modules. Provided technical support and guidance to team members, facilitating smooth workflow and timely completion of projects.
  • Successfully implemented machine learning algorithms that significantly improved the accuracy of resource usage forecasting for the SAKKU Platform.
  • Streamlined data collection process by developing efficient Restful-API modules, resulting in enhanced productivity and reduced time constraints.
  • Enhanced user experience by implementing a file search feature, garnering positive feedback from clients and users.
Publications
Hydra-Mask-RCNN An Adaptive HydraNet Architecture for Unmanned Aerial Vehicle Object Detection
Sara Naseri Golestani; Mahdi SadeghiBakhi; King Fai Ma; Henry Leung
IEEE Open Journal of Instrumentation and Measurement
IEEE 2024
Robust Real-time Magnetic-based Object Localization to Sensor’s Fault Using Recurrent Neural Network
Sara Naseri Golestani, Hamed Rafiei, Mohammad-R Akbarzadeh-T and Alireza Akbarzadeh-T
International Conference on Computer and Knowledge Engineering - ICCKE 2020
ICCKE 2020
Magnetic-based Localization Using Artificial Neural Networks.
Sara Naseri Golestani, H. Rafiei, A. Samadi, H. Hafiz, A. NaddafShargh, M. Akbarzadeh-T,and A. Akbarzadeh.
The 7th joint Congress of Fuzzy and Intelligent Systems 2019
CFIS 2018
Open Source

Most of my work is available in my Github repositories.

  • Furniture Classifier API: An example of how to build a REST API for image classification using PyTorch and Flask.
  • TResnet_ml: Official Pytorch Implementation of "TResNet: High-Performance GPU-Dedicated Architecture" (WACV 2021).
  • Trudeau Posts: This application crawls CNN and Twitter to find the most recent posts of J. Trudeau and displays them.

Skills
Programming languages
Python
I am experianced in coding with python language and been using it for around 5 years.
SQL
I have been Using SQL sine I was working as an AI spacialist, I have enough knoweledge to fetch the data that I need
Matlab
Before Python, I used to use matlab's NN models for faster results
Java
developed a file search module on POD at Fanap.
Frameworks & Tools
Pytorch
My go to library for building NNs from scratch
Tensorflow
If I have to, I'll work with it!
Numpy
What data structure is better than arrays while working with neurons!
Pandas
With this, there's no need of SQL!
Scikit-Learn
A go to library for less computationally expensive ML models
Azure
A cloud platform that offers all sorts of APIs and pre-trained ML models
Flask
A simple nice API handler
Git
How are we gonna keep track of our codes without it?!
Docker
when it comes to versioning our models, we go to docker^^
Certificates
Azure AI Fundamentals (AI900),
Offered by Microsoft
Natural Language Processing
Passed Through Coursera
Python
LinkedIn Badge
Python
Offered through Kaggle
Machine Learning
LinkedIn Badge
Background
University of Calgary
Sep. 2020 - Apr. 2023 | M.Sc. in Electrical and Software Engineering
ENEL 645 Data Mining and Machine Learning A+
ENEL 641 Optimization for Engineers A
ENEL 631 System Identification and Parameter Estimation A
Ferdowsi University of Mashhad
Sep. 2014 - Sep. 2019 | B.Sc. in Electrical Engineering
ENEL FUM Computational Intelligence A
ENEL FUM Micro Processors A+
ENEL FUM Computer Architecture A
ENEL FUM Thesis: Object tracking in robot hand sensing project using magnetic sensors A+
Teaching Assistants
University of Calgary
ENEL 353 Digital Circuits FAL. 2021 and 2022
ENEL 327 Signals and Transforms WIN. 2021 and 2022 and 2023
Ferdowsi University of Mashhad
ENEL FUM Fundamentals of Electronics 2018-2019
Research

I'm interested in various flavors of Machine Learning

  • Machine Learning Engineering,
  • Computer Visions,
  • Machine Learning Development,
  • Statistical Modeling,
  • Deep Learning,
  • Data Science

My principal research interests are in the area of Machine Learning Engineering, Data Science, and Computer Visions.

Throughout my Bachelor study I did not have enough time to explore all sides of data, even though I did gain a lot of experience working in Cognitive Computing Lab (CCL) on a Robot Hand Sensing project to track a magnet inside an amputet hands muscles. This Tracking was done with the help of Recurrent Neural Networks(RNNs) and I was able to publish my very first paper on it. my main contributions during my bachelors project are:

  • ObjectLocalizationinHandinordertoControla Robot Hand Using RNN.
  • Noisecancellationinmagnetic-basedlocalizationusingcomplex-valued neural networks.
  • Robustmagnet-basedlocalizationsystemtosensordis-placementusingthe Genetic Algorithm (GA).

    • During my Master's, I achieved a lot in the field of computer visions with my role as a Research Assistance. I was able to work on various types of images including and not limited to, IR, RBG, UAV and etc. My main contributions could be pointed out as follows:

  • Object detection with the help of Multitask Learning (MTL) to improve the Mask R-CNNs ability to detect small objects with higher precision.
  • This method was applied on backbone of the object detection methods to enrich their output feature maps
  • The Object detection methods used for this improvement on UAV data were YOLO, Mask R-CNN, and DETR.