Leonardo Schettini, Developer in Vienna, Austria
Leonardo is available for hire
Hire Leonardo

Leonardo Schettini

Verified Expert  in Engineering

Artificial Intelligence (AI) Developer

Location
Vienna, Austria
Toptal Member Since
February 22, 2022

Leonardo is a skilled data scientist and software engineer specializing in natural language processing. He has led projects for startups and corporates with varying team sizes. His projects are often written in Python and deployed on Kubernetes and Docker environments. Leonardo is calm under pressure and learns and applies new skills quickly.

Portfolio

Crayon
Agile Software Development, Artificial Intelligence (AI), BERT...
TheVentury
Python, PyTorch, Scikit-learn, SpaCy, Pandas, NumPy, Matplotlib, Time Series...
Instituto Avançado de Tecnologia e Inovação (IATI)
Python, Research, Time Series, Tableau, Oracle, Redmine, Data Processing...

Experience

Availability

Part-time

Preferred Environment

MacOS, Docker, Agile Software Development, Visual Studio Code (VS Code), Python

The most amazing...

...project I've worked on is a series of machine learning tools to make text and audio accessible to hearing-impaired people.

Work Experience

NLP Data Scientist

2022 - PRESENT
Crayon
  • Led development of end-to-end NLP systems for customers around the globe.
  • Leveraged GPT models from OpenAI, Azure services, LangChain, and in-house tools to build a framework for “ask your data” solutions that allowed over delivery of features for customers in 50% of the original time.
  • Researched and experimented with evaluation techniques for the different stages of "ask your data" solutions.
  • Fostered cross-team collaboration by identifying and helping modularize and abstract reusable components.
  • Improved and maintained an internal tool to automate the MLOps lifecycle, using tools like GitHub Actions, Terraform, Docker, and Azure Machine Learning.
Technologies: Agile Software Development, Artificial Intelligence (AI), BERT, Machine Learning, Azure Cognitive Services, Azure ML Studio, Azure SQL Databases, Azure Functions, Terraform, Python, Natural Language Processing (NLP), Generative Pre-trained Transformers (GPT), Named-entity Recognition (NER), Information Extraction, Clustering, Data Processing, Text Processing, Sentiment Analysis, Text Classification, Docker, PyTorch, Pandas, NumPy, GitHub Actions, Azure DevOps, Jira, SQL, Git

Software Engineer | Data Scientist | DevOps Engineer

2019 - PRESENT
TheVentury
  • Served as the lead developer on teams of up to five people for several projects, including a core system for a corporate.
  • Developed an end-to-end machine learning system to make the internet more accessible for the hearing impaired, leveraging semantic similarity and time series forecasting techniques.
  • Developed topic modeling and text generation features for a machine learning product that helps video content creators understand the competition and structure their videos.
  • Worked with early-stage startups, consulting on and developing NLP applications and recommendation systems for recruitment, housing recommendations, and legal documents.
  • Reduced the time to deploy a project's staging environment by 50% while guaranteeing replicability, enabling effortless deployments for feature branches, and allowing easy management of the deployments.
  • Automated several Jira workflows to reduce management overhead.
Technologies: Python, PyTorch, Scikit-learn, SpaCy, Pandas, NumPy, Matplotlib, Time Series, Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), Chatbots, BERT, Kubernetes, Rancher, Docker, Docker Compose, Azure Virtual Machines, GitLab CI/CD, GitHub Actions, Jira, Jira REST API, Machine Learning, DevOps, Node.js, JavaScript, Natural Language Understanding (NLU), Information Extraction, Data Science, Supervised Learning, MySQL, PostgreSQL, REST, Flask, Microsoft SQL Server, Natural Language Toolkit (NLTK), Recommendation Systems, Data Processing, Jupyter Notebook, Research, Information Retrieval, Redis, Agile Software Development, Visual Studio Code (VS Code), MacOS, SQL, Git, Neural Networks, Artificial Intelligence (AI)

Graduate Research Assistant

2018 - 2020
Instituto Avançado de Tecnologia e Inovação (IATI)
  • Conducted a research project in conjunction with CPFL, the second-largest non-state-owned group of electric energy generation and distribution companies in Brazil.
  • Performed descriptive analysis of time series data collected by smart meters.
  • Deployed and maintained the applications required for the research, such as Oracle, Redmine, and Tableau.
Technologies: Python, Research, Time Series, Tableau, Oracle, Redmine, Data Processing, Machine Learning, Linux, SQL, Git, Artificial Intelligence (AI)

Artificial Intelligence Researcher | Data Scientist

2018 - 2019
Recrut.ai
  • Developed a scalable and stateless Python tool for candidate screening that reduced up to 85% of recruitment costs for Unilever and Neurotech.
  • Implemented in-house solutions for tokenization, lemmatization, and named entities recognition for multilingual and unstructured text.
  • Implemented algorithms for automatic ranking of candidates based on their CVs.
  • Deployed and maintained the application server on an AWS Elastic Beanstalk worker environment.
Technologies: Python, Scikit-learn, Natural Language Toolkit (NLTK), Matplotlib, Pandas, NumPy, Supervised Learning, Recommendation Systems, Information Extraction, Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), Natural Language Understanding (NLU), PostgreSQL, REST, Flask, Amazon Simple Queue Service (SQS), AWS Elastic Beanstalk, Data Processing, Jupyter Notebook, Data Science, Research, Information Retrieval, Machine Learning, Agile Software Development, Visual Studio Code (VS Code), Linux, Windows, SQL, Git, Neural Networks, Artificial Intelligence (AI), Clustering, Named-entity Recognition (NER), Text Processing, Sentiment Analysis, Text Classification

Chatbot Developer

2017 - 2018
Elife
  • Developed Facebook Messenger bots for companies such as MasterCard, Yamaha, Super Bock, and Opel.
  • Maintained and improved the in-house chatbot framework based on the Naive Bayes classifier, capable of finding a user’s intent, keeping conversational context, and analyzing a user’s sentiment.
  • Developed an in-house software capable of analyzing Instagram’s pictures, searching for information that the marketing team could use to improve its consulting techniques.
  • Integrated applications with third-party APIs, such as Google Cloud Vision, Google Maps, and client-specific APIs.
Technologies: Node.js, JavaScript, Google Vision API, Databases, REST, Heroku, Data Processing, Agile Software Development, Windows, SQL, Git, Artificial Intelligence (AI)

Recrut.ai

https://recrut.ai/
A Python-based application for screening candidates. As the only developer working on the back end of the tool, I implemented algorithms for text processing, information extraction, and candidate ranking. I also defined the software architecture, making sure the application was scalable. The tool reduced recruitment costs by 85% for Unilever and Neurotech.

Ask Your Data Framework

A framework to speed up the development and deployment of Ask Your Data solutions using OpenAI, Azure stack, LangChain, and other related tools. I introduced an evaluation strategy to the framework to quantify improvements made to the prompt and conversation workflow. I also generalized the data ingestion pipeline and conversation workflow, allowing the framework to be used for multiple projects and reducing the time needed to complete them by up to 50%.

TheVentury

http://www.theventury.com
A Python-based tool for generating movement sequences between sign language signs for a 3D avatar. As a data scientist and lead developer, I integrated the software with internal tools to automate the workflow of animators. I also implemented a heuristic to split movement sequences between signs and transitions based on the speed of the hands and models and algorithms to predict and smooth the movement of generated sequences.

I worked on another Python tool for linking words to their definitions based on the context they were employed to. For this project, I had to train a model capable of working on English and German for domain-specific words.
2013 - 2018

Bachelor's Degree in Computer Science

Federal University of Pernambuco - Recife, Pernambuco, Brazil

2016 - 2017

Bachelor's Degree in Computer Science

University of Vienna - Vienna, Austria

JUNE 2016 - PRESENT

Scrum Weekend

Trampolim Academy

JULY 2014 - PRESENT

JavaScript

Centro Integrado de Tecnologia da Informação (CITI)

Libraries/APIs

Scikit-learn, Pandas, NumPy, Matplotlib, Natural Language Toolkit (NLTK), PyTorch, SpaCy, Jira REST API, Node.js, Google Vision API, Azure Cognitive Services

Tools

Docker Compose, GitLab CI/CD, Jira, Git, Named-entity Recognition (NER), Tableau, Redmine, Amazon Simple Queue Service (SQS), Jupyter, Azure ML Studio, Terraform

Languages

Python, SQL, JavaScript

Paradigms

Data Science, REST, Agile Software Development, DevOps, Scrum, Azure DevOps

Platforms

Rancher, Docker, MacOS, Visual Studio Code (VS Code), Kubernetes, Jupyter Notebook, Linux, Windows, Oracle, AWS Elastic Beanstalk, Heroku, Azure Functions, Azure

Frameworks

Flask

Storage

MySQL, PostgreSQL, Databases, Redis, Microsoft SQL Server, Azure SQL Databases

Other

Natural Language Processing (NLP), Machine Learning, Natural Language Understanding (NLU), Supervised Learning, Artificial Intelligence (AI), Generative Pre-trained Transformers (GPT), Time Series, Chatbots, BERT, Information Extraction, Recommendation Systems, Data Processing, Information Retrieval, Neural Networks, Clustering, Text Processing, Sentiment Analysis, Text Classification, Azure Virtual Machines, GitHub Actions, Research, Cryptography, Streaming Data, Software Engineering, Algorithms, OpenAI GPT-3 API, LangChain

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