Federico Rabinovich, Developer in Buenos Aires, Argentina
Federico is available for hire
Hire Federico

Federico Rabinovich

Verified Expert  in Engineering

Data Scientist and Developer

Location
Buenos Aires, Argentina
Toptal Member Since
September 19, 2022

Federico is an IT professional who has served hundreds of clients over the last 15 years, always going the extra mile to deliver superb products. He became passionate about deep learning five years ago and has worked on various AI-related projects, from time series forecasting with LSTM to computer vision and NLP tasks using ConvNets and Transformers. Federico has built models to automate document verification, saving thousands of hours of manual work and preventing fraud for millions.

Portfolio

Konfio
GPT, Generative Pre-trained Transformers (GPT)...
Grupo Alpha Investing
TensorFlow, Python, NumPy, Long Short-term Memory (LSTM), Time Series...
Swiftpark Australia
PHP, JavaScript, MySQL, HTML, OCR, Computer Vision, APIs, Algorithms, XGBoost...

Experience

Availability

Full-time

Preferred Environment

Jupyter Notebook, Google Colaboratory (Colab), Google Docs, Visual Studio

The most amazing...

...model I've built was able to obtain excellent predictions for short-term fluctuations in the Buenos Aires Stock Exchange, improving my client's ROI by 85%.

Work Experience

Data Scientist | Data Specialist

2021 - 2022
Konfio
  • Tackled various NLP-related tasks for different business use cases. Image processing and CV-related jobs for anomaly detection and fraud prevention.
  • Deployed and productionized models over the AWS infrastructure.
  • Performed monitoring, backtesting, and analytics over the productionized models to gather performance metrics and iterate to deploy new versions.
Technologies: Generative Pre-trained Transformers (GPT), GPT, Natural Language Processing (NLP), Computer Vision, TensorFlow, Python, AWS Lambda, Amazon S3 (AWS S3), Data Science, Deep Learning, Machine Learning, Supervised Learning, Unsupervised Learning, OpenCV, Computer Vision Algorithms, Artificial Intelligence (AI), APIs, Data Analytics, Algorithms, Data Visualization, OpenAI, Generative Pre-trained Transformer 3 (GPT-3), Image Processing, Large Language Models (LLMs), Amazon Web Services (AWS), Anomaly Detection, XGBoost, Decision Trees, Supervised Machine Learning, Keras, Python 3, Convolutional Neural Networks (CNN), OCR, MySQL, Programming, Scikit-learn, Pandas, Jupyter Notebook, SQL, NumPy, Text Processing, Image Analysis, Classification Algorithms, AI Design, MVP Design, AI Programming, PyTorch, Natural Language Understanding (NLU), Language Models, Fine-tuning, Data Scientist, Hugging Face, Graphics Processing Unit (GPU), ChatGPT, OpenAI GPT-3 API, Prompt Engineering, OpenAI GPT-4 API, Time Series Analysis, LightGBM, Neural Networks, Generative AI, Transformer Models, Generative Artificial Intelligence (GenAI), Minimum Viable Product (MVP), BERT

Data Scientist

2018 - 2021
Grupo Alpha Investing
  • Worked on several deep learning and machine learning models to create reliable predictors for stock exchange short-term fluctuations.
  • Implemented and productionized LSTM-based networks to make predictions and operate autonomously in the Buenos Aires Stock Exchange market.
  • Achieved an 85% increase in ROI for clients' investments compared to the previous domain-knowledge investment strategies.
Technologies: TensorFlow, Python, NumPy, Long Short-term Memory (LSTM), Time Series, Machine Learning, Artificial Intelligence (AI), APIs, Data Analytics, Algorithms, Forecasting, Data Visualization, Supervised Machine Learning, Keras, Python 3, Computer Vision Algorithms, Programming, Scikit-learn, Pandas, Jupyter Notebook, Google Colaboratory (Colab), Deep Learning, Supervised Learning, Classification Algorithms, AI Design, Data Processing Automation, MVP Design, AI Programming, Fine-tuning, Data Scientist, Time Series Analysis, Neural Networks, Transformer Models, Minimum Viable Product (MVP)

Software Engineer | Machine Learning

2016 - 2018
Swiftpark Australia
  • Designed and developed a web-based system to manage parking automation in Australia.
  • Outlined and supervised the development of a mobile app for parking inspectors using geolocation and computer vision to pick up license plate digits.
  • Assumed the ownership of the entire project, discussing requirements and goals with the business owner.
Technologies: PHP, JavaScript, MySQL, HTML, OCR, Computer Vision, APIs, Algorithms, XGBoost, Keras, Python 3, Computer Vision Algorithms, Programming, Scikit-learn, Pandas, Jupyter Notebook, Machine Learning, SQL, Deep Learning, Supervised Learning, TensorFlow, Classification Algorithms, MVP Design, AI Programming, Minimum Viable Product (MVP)

Full-stack Web Developer

2008 - 2016
PampaWorks
  • Developed custom content management systems (CMS) for hundreds of websites on demand.
  • Assessed customers' needs and converted them into specific technical requirements.
  • Implemented documented requirements, building hundreds of administrable websites.
Technologies: PHP, MySQL, JavaScript, HTML, CSS, Algorithms, Programming, SQL, C#, MVP Design

Alternative Algorithm to Train ConvNets

As part of my thesis project, I experimented with alternative algorithms to train ConvNets. These are based on preliminary research on how the visual cortex of mammals learns to summarize visible data progressively through stacked layers, just like ConvNets.

NLP Based System for Automatic Quality Assessment on Customer Phone Calls

Several natural language processing (NLP) based modules were developed to evaluate the quality of phone calls between the company's agents and customers. Every NLP module was designed to check whether the agent followed specific rules and policies during the conversation. I set up the algorithms and technics, including speech-to-text, fuzzy matching, and transformer-based models (BERT, GPT-3).

Time Series Forecasting Using LSTM Architectures

Created a stock market autonomous operator for short-term positions. Based on LSTM architecture and using both short-term temporal data and long-term indicators. I developed a system to operate in real-time according to the outcome of its predictions.

AI (CV) Predictor for Estimating Sowing Land and Crop Types

Based on Satellite imagery from different frequency bands, this predictor estimates yields for other crops. Various models were developed and tested with different approaches. Among them are a purely statistical image segmentation, an ensemble of decision trees (XGBoost), and others.

ML Predictor for Detecting When a Car Has Already Park

This project aimed to detect whether the event of parking the car had taken place or not in real-time. The model was meant to be deployed in an edge device inside the car, taking its accelerometer's signals as input to the model and making real-time predictions every few seconds.

Languages

Python, Python 3, SQL, JavaScript, C#, Java, PHP, HTML, CSS

Libraries/APIs

NumPy, Pandas, Scikit-learn, OpenCV, Keras, PyTorch, TensorFlow, XGBoost

Paradigms

Data Science, Anomaly Detection

Platforms

Jupyter Notebook, Amazon Web Services (AWS), AWS Lambda

Other

Google Colaboratory (Colab), Machine Learning, Deep Learning, Computer Vision, Natural Language Processing (NLP), Unsupervised Learning, Supervised Learning, Programming, Web Development, Time Series, Convolutional Neural Networks (CNN), Artificial Intelligence (AI), Computer Vision Algorithms, Algorithms, Large Language Models (LLMs), Supervised Machine Learning, Classification Algorithms, AI Design, Data Processing Automation, MVP Design, AI Programming, Deep Neural Networks, NLU, GPT, Generative Pre-trained Transformers (GPT), Language Models, Fine-tuning, Data Scientist, Hugging Face, Predictive Modeling, Time Series Analysis, Neural Networks, Generative AI, Transformer Models, Object Detection, Minimum Viable Product (MVP), BERT, Long Short-term Memory (LSTM), OCR, APIs, Data Analytics, Forecasting, Data Visualization, Image Processing, Decision Trees, Text Processing, Image Analysis, Natural Language Understanding (NLU), Image Generation, Document Parsing, Graphics Processing Unit (GPU), ChatGPT, OpenAI GPT-3 API, Prompt Engineering, OpenAI GPT-4 API, Generative Artificial Intelligence (GenAI), Retrieval-augmented Generation (RAG), Generative Adversarial Networks (GANs), CI/CD Pipelines, OpenAI, Generative Pre-trained Transformer 3 (GPT-3), Text to Speech (TTS)

Tools

Google Docs, Visual Studio, You Only Look Once (YOLO), Amazon SageMaker

Storage

MySQL, Amazon S3 (AWS S3)

Frameworks

LightGBM

2019 - 2023

Master's Degree in Data Science

University of Buenos Aires - Buenos Aires, Argentina

2010 - 2017

Bachelor's Degree in Systems Engineering

Aeronautical University Institute - Cordoba, Argentina

2010 - 2015

Professional Degree in Systems Analysis

Aeronautical University Institute - Cordoba, Argentina

MARCH 2022 - PRESENT

Computer Vision Nanodegree

Udacity

DECEMBER 2021 - PRESENT

Practical Data Science

AWS

JULY 2021 - PRESENT

Building Cloud Computing Solutions at Scale

Duke University

MAY 2021 - PRESENT

Natural Language Processing

DeepLearning.AI

FEBRUARY 2021 - PRESENT

Tensorflow Developer

DeepLearning.AI

DECEMBER 2020 - PRESENT

AI for Medicine

Deep Learning AI

DECEMBER 2020 - PRESENT

Deep Learning

DeepLearning.AI

DECEMBER 2019 - PRESENT

Data Mining and Knowledge Discovery

University of Buenos Aires

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