Matias Aiskovich, Developer in Helsinki, Finland
Matias is available for hire
Hire Matias

Matias Aiskovich

Verified Expert  in Engineering

Bio

Matias is a machine learning engineer who's delivered creative solutions for social impact projects. His past experience includes working at IBM Research as a machine learning engineer (collaborating with IBM's Yorktown Heights research lab), co-founding a startup that develops research-backed cognitive games for the elderly (which was a provider for a Uruguayan government program), and working on several projects that use machine learning to innovate in the healthcare sector.

Portfolio

Toptal Client
Picture Archiving & Communication Systems (PACS), HL7...
IBM
Computer Vision, Generative Pre-trained Transformers (GPT)...
Caretronics
PHP, Ionic, JavaScript, Pandas, Python, Amazon EC2, Amazon S3 (AWS S3)...

Experience

Availability

Part-time

Preferred Environment

Google Cloud, Python, Linux, Google Cloud Platform (GCP)

The most amazing...

...thing I've built is a machine learning model for age regression from 3D brain MRI images, using the model's delta to diagnose neurodegenerative diseases.

Work Experience

Machine Learning Engineer Consultant

2019 - PRESENT
Toptal Client
  • Led the creation of an ML pipeline for the automatic processing of legal documents for the IFF-DuPont RD team; this project included using open-source (Google's T5) and proprietary (GPT-3) LLM.
  • Worked as a principal machine learning engineer, auditing, improving, and developing processes and coaching team members on a computer vision pipeline for object detection and semantic segmentation to detect small defects in car manufacturing plants.
  • Developed SMART on an FHIR app for a healthtech startup to be published in the Epic (EHR vendor) app store.
  • Integrated the Medweb (a telemedicine company) platform with different EHRs using HL7 and FHIR data.
  • Created gradient boosting machine learning models for predicting DNA sequences' manufacture timeline for Strandbase (a biotech startup).
  • Developed NLP (natural language processing) machine learning sentiment classification models and audit end-to-end machine learning pipeline for marketing startups.
  • Built computer vision and NLP models for extracting information from text for a lifestyle app.
Technologies: Picture Archiving & Communication Systems (PACS), HL7, Fast Healthcare Interoperability Resources (FHIR), Mirth Connect, Python, Medical Imaging, OpenEMR, DICOM, Healthcare, SQL, Amazon Web Services (AWS), Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), Computer Vision, PyTorch, BERT, XGBoost, Genomics, Biopython, Biology, Object Detection, Detectron2, Machine Learning Operations (MLOps), Data Versioning, Amazon EC2, Amazon S3 (AWS S3), Semantic Segmentation, Docker, Word2Vec, Data Mining, Data Modeling, Data Reporting, Neural Networks, Deep Neural Networks (DNNs), NumPy, Amazon SageMaker, Image Processing, Artificial Intelligence, Deep Learning, Artificial Neural Networks (ANN), Databases, Relational Databases, Statistics, 3D Image Processing, Transformers, Hugging Face, Linux, OpenAI, Large Language Models (LLMs), SpaCy, OpenCV, LangChain, Epic Electronic Health Records (EHR), Transformer-XL

Machine Learning Research Engineer

2020 - 2022
IBM
  • Conducted machine learning experimentation in a natural language processing project to detect security threats in software packages for IBM Research in collaboration with the IBM TSS team.
  • Developed computer vision models for age regression (predicting age given an MRI image of the brain) and curated a large dataset of brain MRI images as part of my work in the research task for the Exploratory Life Science Sector (a neuroscience team).
  • Co-authored two research papers: "Sparse Depth Completion with Semantic Mesh Deformation Optimization" (depth perception for augmented reality) and "Acoustic Sensing-based Hand Gesture Detection for Wearable Device Interaction."
  • Coached software engineers on machine learning topics, including NLP and computer vision.
Technologies: Computer Vision, Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), Keras, PyTorch, Docker, Kubernetes, BERT, Statistics, Convolutional Neural Networks (CNNs), Image Recognition, Depth Prediction, 3D Reconstruction, Object Tracking, Data Science, Data Analytics, Data Analysis, Machine Learning, Deep Learning, SQL, Flask, Artificial Intelligence, TensorFlow, Medical Imaging, XGBoost, Machine Learning Operations (MLOps), Amazon S3 (AWS S3), NiftyNet, Word2Vec, Data Mining, Data Modeling, Neural Networks, Deep Neural Networks (DNNs), NumPy, Image Processing, Object Detection, Point Clouds, Artificial Neural Networks (ANN), LiDAR, Databases, Relational Databases, 3D Image Processing, Transformers, Linux, Large Language Models (LLMs), SpaCy, OpenCV

Co-founder

2013 - 2021
Caretronics
  • Created an API using Flask to serve a mobile app that I also deployed on AWS.
  • Developed an app chosen to take part in the Uruguayan governmental project, Ibirapita.
  • Made a web platform based on research for improving the quality of life of people with cognitive diseases, which resulted in the publication “Cognitive Stimulation of Autobiographic and Emotional Memory in a Patient with Alzheimer’s Disease.”.
  • Analyzed the data from patient interactions with the app and tracked patients' progress through time.
Technologies: PHP, Ionic, JavaScript, Pandas, Python, Amazon EC2, Amazon S3 (AWS S3), Databases, Relational Databases

Machine Learning Engineer

2019 - 2020
WeVat
  • Developed machine learning computer vision models with TensorFlow to confirm that retailers' receipts in images were compliant with UK legal norms.
  • Solved performance and scalability problems in company databases, improving their schemas and the general architecture.
  • Designed and implemented the dashboard solution for the company, including carrying out the design and implementation of each dashboard and integrating the data sources that the company uses that formerly were not integrated.
  • Built machine learning models with XGBoost (gradient boosting) to predict the company’s volume of customers.
  • Detected anomalies and potential fraud in data, leading to changes in the platform.
  • Created NLP models to detect receipt features and prevent fraud.
Technologies: Google Cloud, TensorFlow, Keras, Pandas, XGBoost, Scikit-learn, Python, Docker, Kubernetes, Statistics, Convolutional Neural Networks (CNNs), OCR, Image Recognition, BigQuery, Google BigQuery, Data Science, PostgreSQL, Data Analytics, Data Analysis, Machine Learning, Deep Learning, SQL, Flask, Artificial Intelligence, ETL, Data Warehousing, Data Engineering, Computer Vision, PyTorch, Machine Learning Operations (MLOps), Amazon S3 (AWS S3), Word2Vec, Data Mining, Data Modeling, Data Reporting, Neural Networks, Deep Neural Networks (DNNs), NumPy, Image Processing, Artificial Neural Networks (ANN), Databases, Relational Databases, Linux, Google Cloud Platform (GCP), OpenCV

Senior Data Engineer

2016 - 2019
Morsum
  • Designed and led the implementation of an inpatient food ordering project for hospitals based on SMART and FHIR to connect with EHRs.
  • Led implementation of ETL into Google Cloud Platform, using Pub/Sub, Google Dataflow (Java SDK), and Google BigQuery.
  • Developed Python APIs to interface between the web and mobile apps and machine learning models.
  • Created machine learning market basket analysis recommendation models for food ordering.
  • Made scripts in Apache Spark to handle the big data for company products. I parallelized NLP-related tasks of matching food ingredients from many different sources using Spark.
Technologies: Natural Language Toolkit (NLTK), Google Cloud, TensorFlow, Pandas, Scikit-learn, Python, Google BigQuery, Apache Spark, Pub/Sub, HL7, Cloud Dataflow, Apache Beam, BigQuery, Data Science, Java, Data Analytics, Data Analysis, Machine Learning, Deep Learning, SQL, Flask, Artificial Intelligence, Amazon Web Services (AWS), ETL, Healthcare, Fast Healthcare Interoperability Resources (FHIR), Data Engineering, Data Warehousing, Spark, Machine Learning Operations (MLOps), Amazon EC2, Amazon S3 (AWS S3), Data Modeling, Data Reporting, R, NumPy, Neural Networks, Artificial Neural Networks (ANN), Databases, Relational Databases, Natural Language Processing (NLP), Generative Pre-trained Transformers (GPT), Statistics, Linux, Google Cloud Platform (GCP)

Developer | System Administrator

2012 - 2016
Gumma SRL
  • Created in-house software for making quotations with custom company requirements and also inserted it in the SugarCRM.
  • Developed an in-house payroll software for construction projects.
  • Led the project involving server virtualization using VMware.
  • Extended an active directory network to regional offices based in Brazil and Uruguay.
Technologies: Windows Server, VMware, SugarCRM, MySQL, Python, PHP, SQL, Databases, Relational Databases, Linux

Help Desk Worker

2012 - 2012
Grupo Estisol
  • Provided technical support to internal users and infrastructure support on a wide array of technology solutions including Zentyal Servers, VMware ESXi, and Active Directory.
Technologies: VMware ESXi, Windows Server

Interview About Recuerdos

https://www.youtube.com/watch?v=DhtjRrXo_Sc
Here, you can watch a video of my interview where I talk about Recuerdos.

Article About Caretronics

http://www.telam.com.ar/notas/201704/185410-software-app-argentina-adultos-mayores-uruguay.html
This is a link to an article about Caretronics, a startup that I co-founded in 2013.

Harvard CS109A | Final Project

https://harvardfinalproject.wordpress.com/
This is my final project for my Harvard Extension CS109A course where I earned my Harvard data science certificate.

App for Training the Memory of Those in the Uruguayan Elderly Population

This is an app for the elderly to train their memory with content specifically geared towards Uruguayan culture. Currently, it has more than 50,000 downloads. It was selected by the Programa Ibirapitá program through which the government of Uruguay delivers 400,000 tablets to retired citizens, free of charge, with the goal of connecting them to modern technology.

Parrot Detector

https://github.com/maiskovich/parrot_2000
A computer-vision, object-detector project for detecting parrots in images. It is based on Faster R-CNN and implemented using Facebook AI Research's Detectron2 (a library for implementing semantic segmentation and detection algorithms).
2023 - 2024

Master's Degree in Life Science Informatics - Bioinformatics and Systems Medicine

University of Helsinki - Helsinki, Finland

2020 - 2022

Master of Science Degree in Data Science

Universidad Austral - Buenos Aires, Argentina

2016 - 2018

Certificate in Data Science

Harvard Extension School - Boston, MA, USA

2012 - 2014

Commercial Pilot License in Aviation

ETAP - Buenos Aires, Argentina

AUGUST 2017 - PRESENT

Data Engineering for Google Cloud Platform

ROI Training

Libraries/APIs

PyTorch, Keras, XGBoost, Scikit-learn, TensorFlow, Pandas, NumPy, SpaCy, OpenCV, Mirth Connect, Natural Language Toolkit (NLTK)

Tools

BigQuery, Apache Beam, Cloud Dataflow, SugarCRM, VMware, Biopython, Amazon SageMaker

Languages

SQL, Python, Java, JavaScript, PHP, R

Frameworks

Flask, Apache Spark, Spark, Ionic

Paradigms

ETL, Fast Healthcare Interoperability Resources (FHIR)

Platforms

Linux, Docker, Amazon EC2, Google Cloud Platform (GCP), Windows Server, Amazon Web Services (AWS), Kubernetes, NiftyNet, Epic Electronic Health Records (EHR)

Storage

MySQL, Databases, Relational Databases, PostgreSQL, Google Cloud, Amazon S3 (AWS S3)

Industry Expertise

Healthcare, Bioinformatics, Transcriptomics

Other

Computer Vision, Artificial Intelligence, Data Science, Convolutional Neural Networks (CNNs), Deep Learning, Data Analytics, Data Analysis, BERT, Google BigQuery, Machine Learning, Image Recognition, Natural Language Processing (NLP), Data Engineering, Data Warehousing, Detectron2, Word2Vec, Data Mining, Data Modeling, Neural Networks, Deep Neural Networks (DNNs), Image Processing, Artificial Neural Networks (ANN), Generative Pre-trained Transformers (GPT), DICOM, HL7, OpenEMR, Association Rule Learning, Statistics, Medical Imaging, OCR, 3D Reconstruction, Depth Prediction, Object Tracking, Pub/Sub, Object Detection, Machine Learning Operations (MLOps), Semantic Segmentation, Data Reporting, 3D Image Processing, Transformers, Hugging Face, OpenAI, Large Language Models (LLMs), Transformer-XL, VMware ESXi, Picture Archiving & Communication Systems (PACS), Biology, Genomics, Facial Recognition, Data Versioning, Point Clouds, LiDAR, Stable Diffusion, Interviews, Startups, LangChain, scRNA-seq, RNA Sequencing, Bioconductor

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.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

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