Metti Paak, Developer in Toronto, ON, Canada
Metti is available for hire
Hire Metti

Metti Paak

Verified Expert  in Engineering

Machine Learning Developer

Toronto, ON, Canada
Toptal Member Since
June 23, 2020

Metti is a data scientist and machine learning/deep learning expert who has extensive experience in software development and mathematical/statistical modeling. He has worked in the aerospace, manufacturing, and healthcare industries developing custom, data-driven predictive software tools. He is proficient in translating business goals into data products and architecting the entire pipeline to the point of delivery. His work has led to multiple patents, publications, and successful fundraising.


HV Vision
Amazon Web Services (AWS), Python, Machine Learning, Startups...
Celegence LLC
Machine Learning, Artificial Intelligence (AI)...
Artificial Intelligence (AI), Convolutional Neural Networks (CNN)...




Preferred Environment

TFS, Git, PyCharm, Jupyter, Visual Studio, Windows, Linux

The most amazing...

...product I've developed is a deep learning based app that predicts cardiac diseases from non-invasively acquired data, leading to two US patents.

Work Experience

Machine Learning Engineer

2021 - PRESENT
HV Vision
  • Developed a ML pipeline for prediction using hyperspectral imaging data.
  • Developed a model to handle small data with a large number of features. Performed feature engineering and model selection that also reflect the physics of the problem.
  • Communicated with diverse stakeholders: farmers, agronomists, hardware specialists, and executive team.
Technologies: Amazon Web Services (AWS), Python, Machine Learning, Startups, Hyperspectral Imaging (HSI)

Machine Learning and AI Expert

2022 - 2022
Celegence LLC
  • Working with the client to identify the problem and proposed solutions. Implemented the MVP—demonstrating the proposed AI solution—and designed and architected an AI-based solution from the point of data acquisition to prediction.
  • Helped the client's infra team to set up the pipeline backing MLOps requirements. Developed question-answering systems for the systematic review of documents.
  • Developed an automatic pipeline for including or excluding a document for the systematic review.
  • Led the team through agile weakly meetings. Helping Executives better position the new AI-based product for a life science industry-related company.
Technologies: Machine Learning, Artificial Intelligence (AI), Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), Python 3, PyTorch, Language Models, Docker, Machine Learning Operations (MLOps)

AI Research

2021 - 2021
  • Performed AI research on the complex and most recent publications and methods in 3D image segmentation from OCT images for eye disease prediction and triage.
  • Performed AI research on chronic kidney disease trajectory prediction using multi-trajectory mixture models.
  • Presented the results in the form of a live seminar and recorded offline video. Provided critical feedback about where this research provides business opportunities and where there's room for more improvements.
Technologies: Artificial Intelligence (AI), Convolutional Neural Networks (CNN), Medical Imaging, Medicine, Machine Learning, Computer Vision

CTO, Entrepreneur

2020 - 2020
Entrepreneur First
  • Worked on an idea and developed a business model, performed product-market fit analysis, customer development, user interviews. Worked on market sizing and partnership development.
  • Developed cloud-based works-like MVP that uses deep learning to perform risk stratification for eye disease using retina images.
  • Led team building and development, and managed outsourcing.
  • Presented and pitched the idea to VCs and government grants.
Technologies: Cloud, Business Planning, Machine Learning, Image Processing, Deep Learning, TensorFlow, Python

Machine Learning and Data Science

2020 - 2020
Orlando Health
  • Developed a data and machine learning pipeline for ingesting massive healthcare data involving financial and discharge records.
  • Developed predictive models for insurance, readmission, and models for clustering and segmentation.
  • Collaborated with a multidisciplinary team; prepared reports and presentations.
  • Used GCP compute engine to handle big data and model development. Used PySpark for statistical analysis.
Technologies: Visualization, Healthcare, TensorFlow, Scikit-learn, Google Cloud ML, Pandas, Python, Spark, SQL

Senior AI Scientist

2018 - 2020
Analytics 4 Life
  • Developed an ML app that predicts cardiac diseases using non-invasively acquired data leading to pending US/international patents.
  • Led a project to develop a protocol for patient data collection and automated quality assessment, reduced retrial efforts, and cost.
  • Developed an ML-based anomaly detection algorithm for ECG and PPG signals: design of experiment, statistical analysis, ML model training and deployment, and automated reports.
  • Improved inference performance of production ML model deployed on AWS by 60%: parallelization and algorithmic improvements.
  • Led projects on feature engineering, model selection, and robustness analysis: data collection, quality assessment, feature engineering, model selection, deployment, and performance monitoring.
  • Developed a deep learning-based app to predict the respiration information and waveform form the cardiac and blood volume data (ECG and PPG): providing a valuable health marker for disease prediction.
  • Established effective communication with other teams (systems, QA, hardware, clinical), development in close liaison with cardiologists, clinical staff, and end-users.
  • Prepared detailed technical reports for FDA submission; prepared marketing presentations; presented results to executives, investors, and partners; and authored the scientific manuscript.
  • Conducted code reviews, involved in the hiring process, and trained junior data scientists and interns.
Technologies: Amazon Web Services (AWS), Docker, Keras, TensorFlow, MATLAB, Python, C++

Senior Research Scientist, HPC Software Developer

2015 - 2018
Hexagon Manufacturing Intelligence
  • Developed ML models to replace heavy simulations for the design and optimization of auto parts, a fast tool for the initial design and prototyping phase.
  • Developed a high performance parallel linear system solver leading to 1,000 times speed-up: sparse direct solvers and iterative solvers (PCG, Algebraic multigrid).
  • Developed a sparse/dense linear algebra library in C++, scientific code design, and optimization.
  • Developed a tool for the probabilistic analysis of uncertainty propagation from input materials to the final product.
  • Created mathematical, statistical, and numerical models for the simulation of sheet metal processes: differential equations, finite element method, and visualization.
  • Authored, published, and presented a paper on probabilistic analysis of spring back phenomenon in the Institute of Physics (IOP) conference series.
  • Utilized supervised and clustering ML techniques, design of experiments, Monte-Carlo simulation, and HPC.
Technologies: TFS, GPGPU, MPI, OpenMP, UML, Python, C++

Research Scientist

2014 - 2015
Alstom Power
  • Developed a semi-analytical software tool and pipeline for improving the performance of old hydraulic turbines, thereby avoiding significant replacement cost.
  • Developed a tool for predicting what parameters and loads (fluid dynamic forces and conditions) lead to a dangerous structural response (e.h. catastrophe vibrations).
  • Designed and implemented an experimental setup for validating the results of the simulation tool.
  • Prepared documentation and design guidelines, presented results, and published internal white paper.
Technologies: Models, Finite Element Method (FEM), MSC Nastran, ANSYS, MATLAB, Python

Research Scientist

2013 - 2014
Bombardier Aerospace
  • Developed a fast predictive software tool for designing fuselage components with lower weights and yet stable under the loads, reducing the initial design and prototyping cost.
  • Developed design guidelines and design charts for various loads and geometries, resulting in a 10% weight reduction in the initial phase of the design.
  • Developed mathematical and numerical models for the post-instability behavior of the fuselage component.
  • Supervised and trained applications engineer on how to utilize the developed tool and guideline.
Technologies: Finite Element Analysis (FEA), SOLIDWORKS, ANSYS, Python, MATLAB

PhD Researcher

2008 - 2013
McGill University
  • Studied and developed mathematical models for predicting the behavior of complex nonlinear systems: fluid-structure interaction, cardiovascular fibrillation, weather forecast, etc.
  • Developed a high-performance numerical software tool for solving the nonlinear differential equations of the complex systems (airplane wing).
  • Developed a pipeline to compile and deploy the program on a super-computing cluster (Mammoth), request shared (OpenMP) and distributed memory (MPI) resources, and collecting and analyzing Terabyte results.
  • Designed and performed experiments on the nonlinear and chaotic response of elastic shell conveying fluid (artery conveying blood), and validated the results of the software tool.
  • Developed machine learning meta models for forecasting the behavior of the complex systems without running heavy simulations.
  • Taught courses on numerical methods, computer programming, mathematical modeling. Published and presented results in various journals and conferences.
  • Utilized C++, Fortran, Python, distributed computing, OpenMP, MPI, statistical data analysis, stochastic signal processing, and experimental design.
Technologies: Bash, Linux, MPI, OpenMP, Python, Fortran, C++

Predicting Respiration Information and Waveform From ECG and PPG Data

Respiration is an important biological marker for determining the state of health or for predicting an aberration or disease. It is not easy to measure respiration outside clinical settings as it requires inconvenient equipment and sensor attachments. Therefore, it is important to accurately estimate the respiration information from available data such as the ECG and PPG data. I conducted this project in two phases: (i) developing and testing against available data with ground truth and (ii) improving and modifying the program to act on unlabeled data. I completed the project using a variety of tools (data ingestion, feature engineering, deep/transfer learning, ensemble learning) which achieved the desired business target. In phase 2, I modified the model to use propriety data, I implemented a performance monitoring tool to assess potential model drift. This model can be used interactively and in a production environment (deployed on AWS).
In this two-phase project, I developed an object-oriented program to predict RR from ECG and PPG (single or multi-channel). The program ingests data from various sources and in different formats after data preprocessing (anomaly detection, outlier removal, imputation, and quality assessment).

MVP For Eye Disease (glaucoma) Prediction Using Retina Images

For a startup, I developed an AI-based MVP hosted on GCP.

Use case: a retina image is uploaded, image segmentation identifies retina blood vessels, optic cup, and disk (UNet based). This is a multimodal system that also takes into account the eye inter-ocular pressure IOP. The model outputs are (i) a risk stratification score, (ii) patient monitoring longitudinal analysis, and (iii) eye-level information to ophthalmologists as a decision support tool.

Healthcare Startup AI Product Development

I have performed a product-market fit (PMF) analysis, customer development, and user interviews to assess the feasibility and market size for a novel AI-based solution for disease prediction. I provided a pitch deck and presented the idea to venture partners and angel investors. The solution was designed to serve as a platform with an easy to use hardware component to evaluate patients at home or the point of care. I prepared an FDA road map, competitive landscape, and established partnerships.
2008 - 2013

PhD in Computational Mechanics and Complex Systems

McGill University - Montreal, Quebec, Canada


Natural Language Processing


Professional Engineer (PEng)

Professional Engineers of Ontario


Pandas, Scikit-learn, PySpark, TensorFlow Deep Learning Library (TFLearn), REST APIs, TensorFlow, Keras, OpenMP, MPI, PyTorch


MATLAB, Mathematica, Apache Impala, Visual Studio, Jupyter, PyCharm, Git, TFS, SOLIDWORKS


Spark, Flask


Python, C++, Fortran, SQL, UML, Bash, JavaScript, Python 3


Data Science, Parallel Programming, Agile Software Development, RESTful Development, GPGPU

Industry Expertise



NVIDIA CUDA, Linux, Docker, Amazon Web Services (AWS), Google Cloud Platform (GCP)


Apache Hive


Machine Learning, Deep Learning, Mathematical Modeling, Software Development, Time Series Analysis, Natural Language Processing (NLP), Generative Pre-trained Transformers (GPT), Statistics, Signal Processing, Stochastic Modeling, Forecasting, ANSYS, MSC Nastran, Finite Element Method (FEM), Models, Finite Element Analysis (FEA), Computer Vision, Cloud, Big Data, Statistical Modeling, Business Planning, Google Cloud ML, Visualization, Image Processing, Artificial Intelligence (AI), Convolutional Neural Networks (CNN), Medical Imaging, Startups, Hyperspectral Imaging (HSI), Language Models, Machine Learning Operations (MLOps), Medicine

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.


Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring