Hello, I'm
Darwin Saire
I am a Machine Learning Engineer, Data Scientist, NLP Developer, and University Professor.
I hold a Ph.D. in Computer Science from the University of Campinas (UNICAMP), Brazil,
earned with the Recod.ai research lab at the Institute of Computing (IC),
including a research internship at LORIA/INRIA, France.
I am currently expanding my education with an MBA in Software Engineering at USP/ESALQ and a postgraduate
specialization in Applied AI Engineering (LLMs, Generative AI, AI Agents, RAG, MCP, Fine-Tuning & Security).
My skills include strong reasoning, adaptability, and problem-solving, with hands-on experience in the
main machine learning frameworks (e.g., PyTorch, TensorFlow).
I currently work modeling, training, and interpreting NLP models for SiDi - Samsung, located in Campinas - Brazil, developing the Bixby application for mobile, TVs, watches, and fridge devices. Alongside this, I teach part-time as a university professor.
About Me
I'm Darwin Saire Pilco
I have always been deeply interested in statistics, analytics,
and mathematics, which led me to study data science.
I graduated with a bachelor's degree in Systems Engineering (Computer Science)
at the National University of San Agustin (UNSA), Peru.
I followed this with a master's degree in Computer Science
at the University of Campinas (UNICAMP), Brazil.
I completed my Ph.D. at UNICAMP, with a research internship
at the University of Lorraine, France.
More recently, I completed an MBA in Software Engineering
at the University of São Paulo (USP)/ESALQ, Brazil.
My research focused on semantic segmentation tasks and discovered a learning representation
(i.e., context embedding) with a helpful structure and behavior to address the loss of spatial
precision problem present in popular deep learning models.
Specifically, I model the latent space (i.e., give a structure) through Gaussian mixing,
providing a clustering behavior to the features,
which positively impacts the final result of the semantic segmentation.
Interests
Machine Learning.
Deep Learning.
Natural Language Processing.
Pattern Recognition.
Computer Vision.
Generative AI.
LLM.
Language
Spanish (native);
Portuguese (proficient).
English (intermediate).
Education
2026 - Present
Postgrad. Specialization in Applied AI Engineering
Studying LLMs, Generative AI, AI Agents, RAG, MCP, Fine-Tuning, and Security, at UNIPDS/Anhanguera, Brazil.
2024 - 2026
MBA in Software Engineering
Completed at the University of São Paulo (USP)/ESALQ, Brazil.
2017 - 2022
Ph.D. in Computer Science
Completed with GPA 3.85/4.0, at the University of Campinas (UNICAMP), Brazil.
2019 - 2020
Ph.D. internship at LORIA Institute
Ph.D. internship in partnership with INRIA, at the University of Lorraine, France.
2015 - 2016
M.Sc. in Computer Science
Completed with GPA 3.85/4.0, at the University of Campinas (UNICAMP), Brazil.
2008 - 2013
B.Sc. in Computer Science
Completed at the National University of San Agustin (UNSA), Peru.
Experience
Aug. 2026 - Present
Professor - Part-time
Machine Learning instructor and IT educational mentor, teaching statistics,
ML algorithms, and ML libraries.
SENAI/SC.
Aug. 2024 - Jun. 2026
Professor - Part-time
Professor of Statistics and Business Intelligence (BI), using Power BI, Jupyter,
Pandas, Scikit-Learn, and Python. See the Teaching section for details.
FACENS University.
Apr. 2022 - Present
Machine Learning, NLP Developer
Model development for classification in Samsung's virtual assistant (Bixby),
for mobile, watch, TV, and fridge devices.
SiDi - Samsung Institute.
Aug. 2016 - Feb. 2017
Software Consultant
Developed automatic tests to verify correct communication between
Salesforce and JIRA, improving the efficiency in ~ 80%.
zAgile Inc.
Aug. 2013 - Jan. 2015
Researcher - Developer
Implemented GLCM, Gabor, and LBP algorithms for feature extraction and built median and gaussian filters for generating candidate regions, achieving ~ 99% of sensitivity.
Implemented and evaluated CIEL*a*b* algorithm and color-matching functions,
in quality control process, achieving mAP ~ 87%.
Incalpaca TPX S.A
Honors & Awards
First place in the application to the Ph.D. program.
Teaching Assistant, Competitive Programming Course.
Participant, ACM-ICPC South America/South Regional programming contest.
Skills
Teaching
Alongside my industry work, I have taught part-time as a university professor, helping Systems Analysis and Development students build practical, hands-on skills in statistics and data-driven decision-making.
Aug. 2024 - Jun. 2026
Statistics
Taught the Statistics course for the Systems Analysis and Development program:
descriptive statistics (central tendency and dispersion), data visualization
(histograms, scatter plots, boxplots), linear regression, probability fundamentals,
random variables and probability distributions, the normal distribution, sampling,
and estimation with confidence intervals — combining theory with practical
exercises and case studies.
FACENS University.
Aug. 2024 - Jun. 2026
Business Intelligence
Taught the Business Intelligence course, guiding students from data fundamentals
to building real analytics solutions: data warehousing and dimensional modeling,
ETL pipelines, exploratory data analysis and statistics with Python, interactive
dashboards with Power BI, data governance, and an introduction to machine learning
and predictive analytics applied to business decisions.
FACENS University.
Aug. 2026 - Present
Machine Learning Mentorship
Mentoring students as an IT educational instructor, covering statistics,
core machine learning algorithms, and hands-on use of ML libraries.
SENAI/SC.
Certificates
Natural Language Processing in TensorFlow
DeepLearning.AI and Coursera
Issued Dec 2023 - No Expiration Date
Credential ID XH9YJWPC3DGG
Browser-based Models with TensorFlow.js
Coursera Course Certificates
Issued Apr 2023 - No Expiration Date
Credential ID JFQYSDSEYW2H
Generative Adversarial Networks (GANs) Specialization
DeepLearning.AI and Coursera
Issued Feb 2023 - No Expiration Date
Credential ID 3JP4XS4BQLB8
Apply Generative Adversarial Networks (GANs)
DeepLearning.AI and Coursera
Issued Feb 2023 - No Expiration Date
Credential ID JRETTJHW3M3J
Build Better Generative Adversarial Networks (GANs)
DeepLearning.AI and Coursera
Issued Dec 2022 - No Expiration Date
Credential ID P3FDAZN67FXJ
Build Basic Generative Adversarial Networks (GANs)
DeepLearning.AI and Coursera
Issued Oct 2022 - No Expiration Date
Credential ID DEUWK2HZ23SJ
Optimizing Machine Learning Performance
Coursera Course Certificates
Issued Sep 2022 - No Expiration Date
Credential ID 8WYV6TFT6XRZ
Natural Language Processing Specialization
DeepLearning.AI and Coursera
Issued Aug 2021 - No Expiration Date
Credential ID 8XMEX9AHESPX
Natural Language Processing with Sequence Models
DeepLearning.AI and Coursera
Issued Aug 2021 - No Expiration Date
Credential ID ZDV8R8QE3J5T
Natural Language Processing with Attention Models
DeepLearning.AI and Coursera
Issued Jul 2021 - No Expiration Date
Credential ID AP47Z5LZCRNV
Natural Language Processing with Probabilistic Models
DeepLearning.AI and Coursera
Issued Jul 2021 - No Expiration Date
Credential ID VFE7Y9D4RDQ4
Natural Language Processing with Classification and Vector Spaces
DeepLearning.AI and Coursera
Issued Jun 2021 - No Expiration Date
Credential ID H5L73XTDZZQG
Bayesian Methods for Machine Learning (with Honors)
Higher School of Economics and Coursera
Issued Aug 2018 - No Expiration Date
Credential ID A97CKSJLJBS9
Deep Learning Specialization
DeepLearning.AI and Coursera
Issued May 2018 - No Expiration Date
Credential ID VQ3XYBPMFRS2
Convolutional Neural Networks
DeepLearning.AI and Coursera
Issued May 2018 - No Expiration Date
Credential ID EGJFPMTA5QMK
Sequence Models
DeepLearning.AI and Coursera
Issued May 2018 - No Expiration Date
Credential ID GCDK8UE3LPN5
Structuring Machine Learning Projects
DeepLearning.AI and Coursera
Issued Apr 2018 - No Expiration Date
Credential ID E3EDD7BH6EWZ
Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization
DeepLearning.AI and Coursera
Issued Mar 2018 - No Expiration Date
Credential ID UBG5Q9B37GTP
Neural Networks and Deep Learning
DeepLearning.AI and Coursera
Issued Feb 2018 - No Expiration Date
Credential ID CRA92D8T4BD8
Neural Networks for Machine Learning
University of Toronto and Coursera
Issued Feb 2018 - No Expiration Date
Credential ID KJ4CY7EWM4L6
Computer Science
Stanford University
Issued Sep 2014 - No Expiration Date
Credential ID d781a200
Heterogeneous Parallel Programming - CUDA
University of Illinois and Coursera
Issued Mar 2014 - No Expiration Date
Credential ID DMS97ATJ42
Publications
Global and Local Features through Gaussian Mixture Models on Image Semantic Segmentation
Darwin Saire, Adin Ramirez Rivera
IEEE Access - journal, 2022Paper
Code
Graph Neural Blocks on Segmentation
Darwin Saire, Adin Ramirez Rivera
Workshop on International Conference on Computer Vision (ICCV), 2022Paper
Poster
Empirical Study of Multi-Task Hourglass Model for Semantic Segmentation Task
Darwin Saire, Adin Ramirez Rivera
IEEE Access- journal, 2021Paper
Code
Documents Counterfeit Detection Through a Deep Learning Approach
Darwin Saire, Salvatore Tabbone
International Conference on Pattern Recognition (ICPR), 2020Paper
Semantic Segmentation Through Graph Neural Network Blocks
Darwin Saire, Salvatore Tabbone, Adin Ramirez Rivera
Workshop on International Conference on Machine Learning (ICML), 2020Paper
Poster
Semantic Segmentation on Image Using Multi-Task Hourglass Networks
Darwin Saire, Adin Ramirez Rivera
Workshop on Advances in Neural Information Processing Systems (NeurIPS), 2019Paper
Poster
Graph Learning Network: A Structure Learning Algorithm
Darwin Saire, Adin Ramirez Rivera
Workshop on International Conference on Machine Learning (ICML), 2019Paper
Poster
Code
Graph Convolutional Network for Semantic Segmentation Task
Darwin Saire, Adin Ramirez Rivera
Poster on Machine Learning Summer School (MLSS), 2018Paper
Multi-scale Morphological Image Simplification Based on Extrema Relationships: Improvements and Applications
Darwin Saire, Neucimar Jeronimo Leite
Conference on Graphics, Patterns and Images (SIBGRAPI), 2016Paper
Projects
Neural Machine Translation
Neural Machine Translation (NMT) model to translate human readable dates (“25th of June, 2009”) into machine readable dates (“2009-06-25”).
Code
NLP Emoji Generation
The emoji project will help you make your text messages more expressive. So rather than writing “Congratulations on the promotion! Let’s get coffee and talk. Love you!” the emojifier can automatically turn this into “Congratulations on the promotion!. 👍 Let’s get coffee and talk. ☕️ Love you! ❤️”
Code
Jazz Music Generation
A sequence model (LTSM) can be used to generate musical values, which are then post-processed into midi music.
Code
Generate New Dinosaur Names
We have collected a list of all the dinosaur names to create new dinosaur names. We will build a character-level language model to generate new names. Our algorithm will learn the different name patterns and randomly generate new names.
Code
Neural Style Transfer
Neural Style Transfer (NST) is one of the most fun techniques in deep learning. As seen below, it merges two images, namely, a “content” image (C) and a “style” image (S), to create a “generated” image (G). The generated image G combines the “content” of the image C with the “style” of image S.
Code
Face Recognition
Face Verification - “is this the claimed person?” and Face Recognition - “who is this person?”. By comparing two such vectors, you can then determine if two pictures are of the same person using triplet loss function.
Code
Autonomous driving - Car detection
We use object detection using the very powerful YOLO model on a car detection dataset.
Code
Sign language classification
We use a Resnet50 to perform the sign language classification.
Code1
Code2
Deep Neural Network for Image Classification
We use the Convolutional Neural Network (CNN) for cat classification.
Code
Logistic Regression with Neural Network
We build a logistic regression classifier to recognize cats.
Code
Parallel SLIC superpixels
Parallelization of SLIC superpixel Algorithm.
Code
Parallel K-means
Parallelization of K-Means Clustering Algorithm.
Code
License Plate Detection
The automatic vehicle license recognition system (ALPR) has four main modules: 1) Image pre-processing, 2) Detection/location of the license plate, 3) Characters segmentation, and 4) Character recognition.
Code
Contact Me
Workspace
Campinas, São Paulo, Brazil