Summary
Overview
Work History
Education
Skills
Awards
Passing AI courses
Patents And Certifications
Publications
Languages
Timeline
Generic

Dr.Mahdieh Khalili

Richmond Hill,Canada

Summary

Solid understanding and highly skilled with more than 13 years of experience in designing and implementing complex Deep Neural Network models, including transformers, attention mechanisms, Graph Neural Networks, Multi-modal learning, Generative AI, and Large Language Models (LLM). Proficient in fine-tuning and working with large scale dataset and pretrained models, as well as Machine learning techniques such as managing Docker containers and libraries for model deployment. Google Cloud certified and Tensorflow Machine Learning Professional Certified. Having expertise in different Google Cloud services, including Vertex AI, Kubernetes, Big Query, Conversational Platform, NLP platform, and Vision API. Lead and manage ML Research team, provide technical leadership and able to help the team to grow by transferring the knowledge. Filling Patent Application and having more than 16 published papers in Journals and international conferences. Python level programming Expert and good knowlegde of Data Structures and System Design.

Experienced with developing and deploying machine learning models to solve real-world problems. Utilizes strong coding skills and knowledge of neural networks to enhance software performance. Track record of effective collaboration and continuous learning in fast-paced settings.

Overview

9
9
years of professional experience

Work History

Senior AI Scientist

Kinectrics
01.2024 - Current
  • Having hands on different stage of AI projects, from problem definition, define solution, create a pipeline, designing the system, implementing the project until deploying project.
  • Implementing several AI project on Cloud environments (Azure, AWS, GCP).
  • Implementing chatbot for instructed and unstructured data.
  • Having handson working with latest pretrained LLM models such as LLAMA3, Gemini, OpenAI.
  • Fine tune LLMs and Multimodal models with implementing parallel computing.
  • Leading the team.
  • Collaborate with world class international team.

AI Professor

George Brown College
01.2024 - Current
  • Teaching the course of Applied Mathematics of Deep learning and Advanced Mathematics of Deep learning.
  • Designed the materials of the course with latest AI algorithm and teaching the course.
  • Help to the students to define their capstone project in AI.
  • Guide students for contributing with competition with Kaggle.
  • Able to guide and manage international students.

AI Professor

Centennial College
09.2024 - Current
  • Teaching the course of Deep learning and machine learning.
  • Manage the international class.
  • Guide students to define the capstone projects.
  • Able to teach all the Machine learning and deep learning algorithms to the students.
  • Perform tests and quiz from students and provide solution.

Senior Research AI Scientist

APPLIED RECOGNITION
04.2023 - Current
  • Design and implement cutting edge ML research line, create POC of the project, Leading the team members, implement, train and optimized solution for large scale dataset of the face, Train the model on GCP with V100 Accelerators and A100 local GPUs
  • Collaborate closely with business partners, product and different engineering teams
  • Project: GENERATIVE AI: 3D face Generation with Using 2D data of face data (Work in Progress)

Senior AI Research Scientist

QUANTIPHI-CANADA
01.2022 - 03.2023
  • Designing and technical leading the project for 'Multi-modal Molecule Match': Lead and manage world class research and engineering teams
  • Design and implement cutting-edge Multi-modal ML models with large scale dataset with using NVIDIA GPUs and SDKs
  • Create foundation model based on multi-modal learning (text and molecule structure(Structures-Text)) in self-supervised learning Manner with implementing Custom model with Transformers and contrastive loss
  • Define downstream tasks on top of model foundation model, such as Information Retrieval task and generate molecule, based on scientific text description of the molecule
  • Build automated mechanisms for evaluating, measuring, and deploying the algorithms and/or models that was developed
  • Build monitors for algorithm or model performance and make recommendations for adjustments to increase accuracy
  • Technical work: Using NVIDIA pretrained Image(MegamolBart) and SciBert for create embedding and use joint representation layers to bring two different modalities (Text description and Molecule Representation) together with contrastive loss
  • Create the best result based on other works that exist
  • Define and lead research line in multi-modal and Generative AI, talking and presenting work regularly with other stack holders of the project
  • Having hands on coding and check the quality of the codes of the team, doing daily research for publication and filled one patent application
  • Projects: MULTIMODAL GENERATIVE AI: Visual Question Answering with open, closed answers: Working with several Multi-modal Pre-trained models in Text and Vision modalities (such as Gato, Flava, Visual Bert, CLIP, DALLE)
  • Working with world class level ML research and engineering team
  • Create Visual Question Answering system for Medical images for closed and opened answers
  • Get the best results respect to other research work
  • Regularly present the work to stakeholders and other business partners
  • Performing Research and ML engineering work in Google Cloud Platform

Senior Machine Learning Engineer

QUANTIPHI- CANADA
02.2021 - 01.2022
  • Participating, mentoring and having hands on in coding in several NLP projects implemented in Google Cloud For Question Answering and Content Search in the text, and Text summarization in Medical Text Documents
  • Participating, mentoring and having hands on coding in several Computer Vision projects such as: Anomaly Detection in Image (Detecting accident area of the car for Insurance company) Image Captioning from Molecule Image to Text(InChI), participating in Kaggle competition and get position and arrive to a contract with host of the project
  • Layout Analysis of the documents such as Text detection, Image detection, Table detection in Documents

Machine Learning/ Deep Learning Specialist

AGS COMPANY
04.2016 - 01.2020
  • Company Overview: Working with world class level Engineering Team members and clients
  • Designing and Teaching Tailored ML and DL courses for clients based on the needs of the clients and famous universities in the Europe with more than 600 hours
  • Collaborating with research and Engineering team to define the guidelines
  • Finding Solution for AI Business problem in OCR, Image Processing and Computer Vision
  • Having experience on these challenges in Computer Vision:Object Detection Algorithms (such as R-CNN, Fast-R-CNN, Faster R-CNN), Image Segmentation (Mask-RCNN), Instance Segmentation, Object Detection and Localization, Training Customize Data set with YOLO, Pose-Estimation, Real-Time Face Detection, Land-Mark Detection
  • Participating and having hands-on in coding, testing and documenting of these Projects: Trading Market Prediction with Times serious data, Natural Language Processing: Text summarization, OCR: Detecting ID cards information in scan documents(with using convolution neural network and lstm), Writing software for evaluating the quality of labelled data that was labelled by labellers group
  • Working with world class level Engineering Team members and clients
  • Used technologies: Working with Linux environment (Ubuntu), Having experience and challenge with these deep learning algorithms and topics: implementing MLP, CONV, Auto-encoder and LSTM in Keras and Tensorflow, GAN, Transfer Learning (VGG, ResNet, Inception V3, MobileNetV2), Inception Network, Ensemble Learning, Scikit-learn Pipelines
  • Implementing deep learning algorithms in GPUs
  • Implementing code environment with Git repositories, and Docker, working with Jupyter notebook and Pycharm
  • Teaching Machine learning and Deep Learning course to the clients and team

Machine Learning Engineer

APPLIED RECOGNITION
06.2017 - 11.2017
  • Generate 2D face data with using GANs for generating synthetic dataset for training face detection algorithm
  • Create face detection pipeline and implementing face detection pipeline

Education

Ph.D. - Computational Intelligence and Machine Learning

UNIVERSITY OF GENOVA
Italy
04-2016

Master - Information Technology Engineering (E-Commerce)

UNIVERSITY OF QOM
01.2011

Bachelor - Applied Mathematics

TEHRAN CENTRAL BRANCH OF AZAD UNIVERSITY
01.2009

Skills

- Multi-modal,

- Generative AI,

- LLM, Computer Vision,

- NLP,

- Image Processing,

- Transformers,

- Attention Mechanism,

- PyTorch,

- Tensorflow,

- Keras framework,

- Gitlab,

- Docker,

- Code testing and documentation(Sphinx),

- Google Cloud,

- AWS,

- ML and AI Libraries: Pandas, Scikit-learn,Numpy, Scipy,PIL, OpenCV,Data Visualization(Seaborn, Matplotlib),

-MLOps,

-Containers, Linux,Unit-test,Distributed Training,

- Experience with NVIDIA Pre-trained models,

-NVIDIA Multi processing with GPUs,SDK

Awards

  • Three years full paid scholarship for Ph.D. from University of Genoa, Italy.
  • Two years full paid scholarship for Master in Information Technology Engineering.

Passing AI courses

Passing AI Specialist courses:

  • Reinforcement Learning in Coursera
  • Natural Language Processing Specialization
  • Generative Adversarial Networks (GANs) Specialization
  • Introduction to Self-driving Cars
  • State Estimation and Localization for Self-Driving Cars
  • Visual Perception for Self-Driving Cars
  • Motion Planning for Self-Driving Cars

Patents And Certifications

System and Method for Translating Image of Structural Formula of Chemical Molecule into Textual Identifier, QUAN013US:28122022, 12/28/22, Tensorflow Developer Certificate, 2021, 56233269, Google Cloud Machine Learning Engineer

Publications

  • Tri-Phasic CT liver Characterization and colour data fusion, Dellepiane. G.S, Khalilinezhad. M, Ferretti. R, International Journal of Advance Research in Computer and Communication Engineering, 4, 08/2015, 2278-1021
  • Detecting HCC Tumor in Three Phasic CT liver Images with Optimization of Neural Network, Khalilinezhad. M, Dellepiane. Vernazza. G, International Journal of Medical, Health, Biomedical, Bioengineering and Pharmaceutical Engineering, 9, 3, 2015
  • A Fuzzy Expert System for Assessing the Internet Stored Based on Web Site Attributes, Khalilinezhad. M, Nadali. A, Dehghani. N, Ghazivakili. M, ACEEE Int. J. on Information Technology, 3, 3, 09/2013
  • Extracting Hidden Patterns in Blood Donor Database using Association rule Mining, Khalilinezhad. M, Dellepiane. S, Abedi. F, Vernazza.G, European Data mining conference, Lisbon, Portugal, 07/2014
  • Innovation Culture Assessment by Fuzzy System (Case Study: An Iranian IT Company), Goldasteh.P, Nadali.A, Khalilinezhad. M, IEEE 14 on Information Management, Innovation Management and Industrial Engineering, 2011, China
  • Prediction of Healthy Blood with Data Mining Classification by using Decision Tree, Naive Bayesian, and SVM Approaches, Khalilinezhad. M, Minae. B, 3rd International Conference on Image, Vision and Computing (ICIVC 2014), Paris, France, 01/2014
  • Evaluating Online Shopping Stores Based on Website Attributes Using Fuzzy Expert System, Khalilinezhad. M, Nadai.A, ACEEE International Conference on Advanced in Computer Science and Application, CSA, 2012
  • Evaluation Discovered Rules from Association Rules Mining Based on Interestingness Measures Using Fuzzy case study: Credit Score of bank customers Expert System, Khalilinezhad. M, Minae. B, IEEE 14th Application of Digital Information and web Technologies, 2012, USA
  • Maturity Assessment of an Information Technology Organization Using Fuzzy Expert System Based on OPM3, Khalilinezhad. M, Nadali. A, IEEE Intentional Conference on Economic Management and Engineering Technology (ICETEM 2012), 2012, China
  • Innovation Culture Assessment by a Fuzzy Expert System, Nadali. A, Khalilinezhad. M, IEEE 14th on Information Management, Innovation Management and Industrial Engineering, 2011, China
  • Probability Prediction of Having HIV, Hepatitis B, C with using Data Mining Patterns in Blood Transfusion Organization, Khalilinezhad. M, Abedi. F, Nadali. A, Minaee.B, First Conference in Information Technology in Health, 2010, Hormozgan, Iran
  • Data mining in Blood Transfusion, Khalilinezhad. M, Minaee. B, 2010, Tehran, Iran
  • Identifying the User Behavior Using Association Rules, Khalilinezhad. M, Minaee. B, 2010, Iran
  • Data mining in Medicine, Khalilinezhad. M, Minaee. B, 2010, Tehran, Iran
  • Strategic planning IS/IT: Approaches & Mechanism, Khalilinezhad. M, Minaee. B, 2010, Tehran, Iran
  • Using data mining to discover hidden patterns in identifying the main cause of exemption in blood transfusion organization donors of Hormozgan province, Khalilinezhad. M, Abedi. F., Minaee. B., 2008

Languages

English
Native or Bilingual
Italian
Full Professional
Persian
Native or Bilingual

Timeline

AI Professor

Centennial College
09.2024 - Current

Senior AI Scientist

Kinectrics
01.2024 - Current

AI Professor

George Brown College
01.2024 - Current

Senior Research AI Scientist

APPLIED RECOGNITION
04.2023 - Current

Senior AI Research Scientist

QUANTIPHI-CANADA
01.2022 - 03.2023

Senior Machine Learning Engineer

QUANTIPHI- CANADA
02.2021 - 01.2022

Machine Learning Engineer

APPLIED RECOGNITION
06.2017 - 11.2017

Machine Learning/ Deep Learning Specialist

AGS COMPANY
04.2016 - 01.2020

Master - Information Technology Engineering (E-Commerce)

UNIVERSITY OF QOM

Bachelor - Applied Mathematics

TEHRAN CENTRAL BRANCH OF AZAD UNIVERSITY

Ph.D. - Computational Intelligence and Machine Learning

UNIVERSITY OF GENOVA
Dr.Mahdieh Khalili