VIEIRA SANTOS

Machine Learning Engineer

Rua Cordovil, Rio de Janeiro - RJ, Brazil [21250-450]

+55-21-394*-****

adzglt@r.postjobfree.com

https://www.linkedin.com/in/vieira-santos-50a878285/ Summary

Innovative Machine Learning Engineer specializing in Deep Learning with a profound understanding of the field, grounded in strong mathematical background. Switching career and 5+ years of experience in diverse DL/ML projects, including Face Recognition, Speech Recognition and NLP, have demonstrated my adaptability to new domains. Consistently upholding a strong moral work ethic, coupled with flexibility and eagerness to learn cutting-edge techniques, has repeatedly contributed to my success throughout my career. I am willing to take new adventures and utilize my professionalism to make a positive impact on the world to make the world a wonderful place. Technical Skills

Programming Languages Python • C/C++ • MATLAB • Java Deep Learning Frameworks Numpy • Scikit-learn • SciPy • Tensorflow • PyTorch • MXNET • Caffe • Keras • Matplotlib • Kaldi

Deep Learning Techniques CNNs • RNNs • Attention Mechanism • GANs • Transfer Learning Cloud Computing Platforms AWS • Google Cloud Platform • Microsoft Azure Career Experience

Inovação Médica Goiânia, Brazil Mar 2023 – Present Machine Learning Engineer, Part-time

Oversaw end-to-end development and deployment of a personalized AI healthcare Assistant.

• Implemented personalized AI healthcare Assistant pipeline using ChatGPT API and LangChain.

• Recommended and supervised fine-tuning of the Portuguese-English translator using medical data, which improved the accuracy by 2.3% in healthcare helpline questions and answers.

• Raised the accuracy of questioning by 3.2% with prompt engineering.

• Tutored 3 interns for a duration of 2 months, giving them the total overview of machine learning project pipeline.

Almawave do Brasil Informática Ltda. São Paulo, Brazil Sep 2019 – Dec 2022 Machine Learning Engineer, Full-time

Upgraded the Text-to-Speech system despite it being a new field to me.

• Enhanced the naturalness by 0.3 on the MOS scale using modern TTS models.

• Added 100 speakers using Transfer Learning, requiring less than 20% of data previously needed training.

Played a key role in developing both online and offline Speech Recognition systems.

• Trained the acoustic model on 23, 000 hours of speech data using Kaldi Framework on Ubuntu environment.

• Reduced the inference time by customizing the model to a smaller size, resulting in a 50% decrease.

• Successfully deployed to the android application, which gained more than 1 million users in 3 months.

• Optimized the language model inference to reduce the memory allocation size by 60% on mobile devices.

UNICAMP Campinas, São Paulo, Brazil Oct 2017 – Sep 2019 Postgraduate Researcher

Participated in various projects as a Deep Learning postgraduate researcher, making substantial contributions.

• Established a face recognition system pipeline including preprocessing, face detection, feature extraction etc.

• Suggested and implemented various methods including center-loss, arc-loss and inception-resnet, resulting in an increase of 1.9% in face verification accuracy.

• Proficiently improved face detection speed by 1.4 times and detected 12 more facial landmarks with greater accuracy using MTCNN methods.

• Integrated an anti-spoofing module based on deep learning, achieving an impressive accuracy of more than 90%.

• Deployed the system to a mobile application by customizing and combining features of 5 smaller models, such as MobileFaceNet v1 and v2 etc.

• Prepared only for 1 month and participated individually in Fake Face Detection Challenge in 2018, taking 6th place.

• Designed and implemented a face dataset refining tool utilizing the aforementioned feature extraction module and clustering algorithms beyond the call of duty, resulting in the removal of 10% of irrelevant face images.

Why Me

• Strong theoretical foundation and practical expertise in ML, especially Deep Learning

• Quick learner and incredibly productive

• Highly responsible and conscientious

Education

• Master of Science in Computer Science, University of Campinas, Campinas Brazil, 2017-2019

• Bachelor of Science in Mathematics, University of Campinas, Campinas Brazil, 2011-2017

Contact this candidate