Thalles Silva
AI Research Scientist at Corteva Agriscience
São Paulo, Brazil
Summary
🤩
Rockstar🎓
Top SchoolThalles Silva is an AI research scientist and PhD candidate at the University of Campinas with 11 years of experience applying machine learning to real-world problems across industry and research. He has a strong track record in representation learning, computer vision, and multimodal approaches, contributing practical implementations of SimCLR and DeepLab_v3 that improve contrastive learning and semantic segmentation pipelines. At Corteva Agriscience he is bringing transformer architectures to genomic (DNA) data, blending biology and large-scale AI methods. His background spans roles from software engineering and R-based scientific tooling to senior data science and AI architecture, reflecting both production-grade engineering and academic rigor. An avid writer on deep learning, he combines open-source contributions and publications to translate research advances into robust, deployable systems.
11 years of coding experience
11 years of employment as a software developer
Machine Learning Nanodegree Machine Learning, Machine Learning Nanodegree Machine Learning at UdacityDoctor of Philosophy - PhD Artificial Intelligence, Doctor of Philosophy - PhD Artificial Intelligence at Universidade Estadual de CampinasBachelor Computer Science Computer Science, Bachelor Computer Science Computer Science at Universidade Estadual de Santa CruzBachelor Computer Science Computer Science, Bachelor Computer Science Computer Science at Algoma University
English, Portuguese
Github Skills (16)
semantic-segmentation10
mask-rcnn10
representation-learning10
faster-rcnn10
pytorch10
machine-learning10
unsupervised-learning10
deep-learning10
tensorflow10
resnet10
python10
contrastive-learning10
data-augmentation9
data-preprocessing9
tensorboard9
Programming languages (6)
Github contributions (5)
sthalles/deeplab_v3
Dec 2017 - Sep 2018
Tensorflow Implementation of the Semantic Segmentation DeepLab_V3 CNN
Role in this project:
ML Engineer
ML EngineerContributions:63 commits, 1 PR, 56 pushes in 9 months
Contributions summary:Thalles contributed to the implementation of the DeepLab_v3 semantic segmentation model using TensorFlow. Their work included adding default parameters, removing unnecessary code, and making improvements to the DeepLab network. They also modified data loading and preprocessing functions, and integrated a ResNet-50 checkpoint for feature extraction. The commits demonstrate the user's involvement in setting up and running the training process for the model.
pythonsemantic-segmentationdeeplabdeep-learningtensorflow
sthalles/SimCLR
Feb 2020 - Feb 2021
PyTorch implementation of SimCLR: A Simple Framework for Contrastive Learning of Visual Representations
Role in this project:ML Engineer
ML EngineerContributions:2 releases, 144 commits, 5 PRs in 1 year
Contributions summary:Thalles implemented and refined a SimCLR model for contrastive learning of visual representations using PyTorch. They added features such as cosine similarity calculation, temperature parameters, L2 normalization, and random Gaussian blur augmentations to the training process, indicating a focus on improving the model's performance and robustness. The commits show modifications to the training loop and loss function calculation.
pytorchsimclrrepresentationssimple-frameworkunsupervised-learning
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.